eBook - Data Visualization with R

Data and R Code Files for Data Visualization with R

Data and R Code

<h3>The Purpose</h3>



<p>The point of a visualization is to communicate its data in a quick and meaningful way while remaining 100 percent accurate. A visualization should serve a clear purpose and not overwhelm the users with unnecessary details. If possible, a visualization should be designed to encourage the users to compare its various elements so as to give insight into the meaning behind the data.</p>



<p>So, to design a data visualization to get its message through, you need to first understand the data itself. Then use well-known design patterns and use the type of visualization that reveals the data in the best way.</p>



<h3>Understand the Data</h3>



<p>To design an effective visualization with a clear purpose, you should thoroughly understand your data. The following points will help you find the information you want to convey with your data:</p>



<ul><li>What kind of data is it? Nominal, ordinal, interval, or ratio data?</li><li>How different parts of the data relate to each other?</li><li>Can you organize the data in a way that makes it easy for you to create your visualizations?</li><li>What do you want to communicate with your data?</li></ul>



<p>It's easy to start thinking about how the visualization should look, but when you have answered these questions it will be easier to decide what kind of visualization you should use and how it will look and communicate its data.</p>



<h3>Use Well-known Design Patterns</h3>



<p>When you have understood the data, how it is organized, and how its parts relate to each other, you should consider using well-known design patterns to communicate your data. For example, if you want to show how a measure behaves over time, you should use a line chart because its strength is that it tells your users a lot without their having to look at the specific details.</p>



<h3>Design Individual Elements to Reveal the Data</h3>



<p>Apart from the design pattern you choose, an effective visualization is also about how you design and make individual data elements stand out and reveal the data. In other words, the design of the individual elements of a visualization should reveal the data to your users in a quick and intuitive way. An easy way of achieving this is to use a different color on one of the dots in a group of many dots. The different color makes it much easier for you to find the dot and reduces your load of information. Other examples of intuitive design are:</p>



<ul><li>Position</li><li>Orientation</li><li>Size</li><li>Shape</li><li>Color hue, brightness and saturation</li></ul>



<h3>Avoid the Pitfalls of Data Visualization</h3>



<p>To experience the benefits of data visualizations you must avoid the pitfalls. Here are some common pitfalls:</p>



<h4>Color abuse</h4>



<p>Do not overdo colors. Be aware that the wrong color in the wrong place might cause confusion rather than clarity.</p>



<h4>Misuse of pie charts</h4>



<p>Avoid having pie charts side by side to compare. Try not to squeeze too much information into them.</p>



<h4>Visual clutter</h4>



<p>Too much information defeats the purpose of clarity. Use a maximum of nine KPIs and remove all visual clutter.</p>



<h4>Poor design</h4>



<p>A beautiful visualization is not necessarily the most effective. Use design best practices at all times.</p>



<h4>Bad data</h4>



<p>Spot and correct issues with your data before you present it. Do not let your visualization take the blame for bad information.</p>



<h3>Commonly Used Visualizations</h3>



<p>In this course, we will study and understand the most commonly used data visualizations.</p>



<ul><li><a href="https://financetrain.com/understanding-bar-chart-column-chart-and-area-chart/">Bar Chart and Column Chart</a></li><li><a href="https://financetrain.com/line-chart-when-to-use-it/">Line Chart</a></li><li><a href="https://financetrain.com/understanding-bar-chart-column-chart-and-area-chart/">Area Chart</a></li><li><a href="https://financetrain.com/what-are-pie-chart-and-donut-chart-and-when-to-use-them/">Pie Chart</a></li><li>Scatter Plot</li><li>Bubble Chart</li><li>Box Chart</li><li>Japanese Candlestick Chart</li><li>OHLC Chart</li><li>Stock Chart</li><li>Treemap</li><li>Heatmap</li><li>Maps</li></ul>

Overview of Data Visualization

<h3 id="lecture_heading">Bar Chart and Column Chart</h3>



<p>A Bar Chart or Bar Graph can be only used to compare values. It presents grouped data using rectangular bars whose lengths are proportional to the values that they represent. The bars can be plotted vertically or horizontally. A vertical bar chart is sometimes called a Column Chart. Bar charts are usually scaled so that all the data can fit on the chart. Bars on the chart can be arranged in any order. The point values can be compared between themselves (single-series chart) or values inside the category (multi-series chart). In case of several series points are grouped by categories.</p>



<p>While talking about charts, we will generally refer to different axis as dimensions and measures.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/bar-chart-1.png" alt="" class="wp-image-26999"/></figure></div>



<p>The same chart can be presented as a column chart as shown below:</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/column-chart-1.png" alt="" class="wp-image-27000"/></figure></div>



<h3 id="lecture_heading">Area Chart</h3>



<p>An area chart or area graph is used to display quantitative data in a graphical manner. Common uses are the comparison of two or more quantities. They can be used to represent cumulated totals using numbers or percentages over time. Use the area chart to show trends over time among related attributes. When multiple attributes are included, the first attribute is plotted as a line with color fill followed by the second attribute, and so on. Technically this chart type is based on Line Chart and represents filled area between zeroline and the line that connects data points. The same as for Line Chart, timeline scale is the most common case for the X-axis of the Area Chart, but sometimes ordinal scale can be also used.<br><br>The following area chart shows venture capital investment by stage.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/27406-chart-1.png" alt="" class="wp-image-27001"/></figure></div>



<h3>100% Stack Area Chart</h3>



<p>100 Percent Stacked Area Charts display the data that can be put in a table on a worksheet (the items should have two coordinates or parameters). A stacked chart means that all values of one category form a whole; when we’ve got a 100% stacked chart, it means that this chart displays the comparison of the percentage value each part of the category brings to the category. Real values are not displayed, chart shows only the percentage.</p>



<p>This 100% stacked area chart shows the venture capital investment by stage.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/27406-chart-2.png" alt="" class="wp-image-27002"/></figure></div>



<p></p>

When to Use Bar Chart, Column Chart, and Area Chart

<p>A Line Chart or Line Graph is the most popular type of the data visualization. It displays information as a series of data points called 'markers' connected by straight line segments. Usually, a Line Chart shows a dynamics of the displayed value in time (in case of Cartesian Chart) or by X (in case of Scatter Chart).</p>



<p>The following line chart shows the Moody's Corporate Bonds yields for Baa and Aaa corporate bonds.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/27393-chart-1.png" alt="" class="wp-image-27006"/></figure></div>



<p>We can also present the same chart as a spline, which consists of a set of points connected with a curve. Notice that the data points are connected with a fitted curve instead of straight lines.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/27393-chart-2.png" alt="" class="wp-image-27007"/></figure></div>



<h3>When to use it</h3>



<p>The line chart is primarily suitable when you want to visualize trends and movements over time, where the dimension values are evenly spaced, such as months, quarters, or fiscal years.</p>



<p>Your data set must consist of at least two data points to draw a line. A data set with a single value is displayed as a point.</p>



<h4>Advantages</h4>



<p>The line chart is easy to understand and gives an instant perception of trends.</p>



<h4>Disadvantages</h4>



<p>Using more than a few lines in a line chart makes the line chart cluttered and hard to interpret. For this reason, avoid using more than two or three measures.</p>

What is Line Chart and When to Use It

<h3>Pie Chart</h3>



<p>The Pie Chart is essentially a circle divided into sectors. The area of each item reflects its value's proportion of the sum of all values in one data set.</p>



<p>Pie Chart are useful when you need to display the share of each constituent part as compared to the whole volume. Sectors can be not only depicted within the whole circle, but also separated from the rest of the chart making it an exploded Pie Chart. This kind of circular graphic remains illustrative only when provided with a few constituent parts. Pie Chart with too many slices are hard to work with efficiently.</p>



<p>The following chart shows the breakup of Full Time Employees by region at HSBC (As per Annual Report 2015)</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/27411-chart-1.png" alt="" class="wp-image-27011"/></figure></div>



<h4>When to use it</h4>



<p>The primary use of a pie chart is to compare a certain sector to the total. The pie chart is particularly useful when there are only two sectors, for example yes/no or queued/finished.</p>



<h3>Donut Chart</h3>



<p>A Doughnut Chart or Doughnut Graph is a variant of the pie chart, with a blank center allowing for additional information about the data as a whole to be included. Each point is specified by an arc that length is proportional to the circumference as the data value to the total sum of all values.</p>



<p>The same chart above can be presented as a donut chart as shown below:</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/27411-chart-2.png" alt="" class="wp-image-27012"/></figure></div>



<h4>Advantages</h4>



<p>The pie chart provides an instant understanding of proportions when few sectors are used as dimensions. When you use 10 sectors, or less, the pie chart keeps its visual efficiency.</p>



<h4>Disadvantages</h4>



<p>It is often hard to compare the results of two pie charts with each other, and therefore you should not do it.</p>



<p>It may be difficult to compare different sectors of a pie chart, especially a chart with many sectors.</p>



<p>The pie chart takes up a lot of space in relation to the values it visualizes.</p>

What are Pie Chart and Donut Chart and When to Use Them

<p>Scatter charts show the relationships among the numeric values in several data series, or between two groups of numbers as one series of XY coordinates.</p>



<p>A scatter chart has two value axes, showing one set of numerical data along the horizontal axis (X-axis) and another one along the vertical axis (Y axis). It combines these values into single data points and displays them in uneven intervals, or clusters. Scatter charts are commonly used for displaying and comparing numeric values, such as scientific, statistical and engineering data.</p>



<p>Scatter plots are best used for data sets in which there is likely to be some form of relationship or association between two different elements included within the data, for example, the relationship between Sales and Quantity per Customer. Financial analysts often use these plots to show correlation between two funds, two stocks or even a stock and the general market.</p>



<h4>Example</h4>



<p>The following chart presents the relationship between salary and age for 474 employees of a company. This type of data would be expected to show some form of trend since, as the staff gains experience, you would expect their value to the company to increase and therefore their salary to also increase.</p>



<p>As an example, one of the employees as 28.5 years old and had a salary of $16,080. To put this data onto a scatter plot, we insert age into the horizontal axis and salary onto the vertical axis. The different entries onto the plot are the 474 combinations of age and salary resulting from a selection of 474 employees, with each observation being a single point on the chart.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/27422-scatter-plot.png" alt="" class="wp-image-27016"/></figure></div>



<p>The charts shows that in fact for this company there is no obvious relation between salary and age. We can observe that the age range of employees is from 23 to 65. We can also see that a lone individual earns a considerably higher salary than all the others and that starters and those nearing retirement are actually on similar salaries.</p>



<h4>Interpretting a Scatter Plot</h4>



<p>Having drawn the plot it is necessary to interpret it. The author should do this before it is passed to any user. The most obvious relationship between the variables X and Y would be a straight line or a linear one. If such a relationship can be clearly demonstrated then it will be of assistance to the reader if this is shown explicitly on the scatter plot. This procedure is known as linear regression and we will discuss it in another course.</p>



<h3>Bubble Chart</h3>



<p>A Bubble Chart or Bubble Graph is used to display three dimensions of the data. Each datapoint is plotted as a disk that expresses two of the values through the disk's xy location and the third through its size. Bubble charts can facilitate the understanding of social, economical, medical, and other scientific relationships. A Bubble Chart can be plotted on both Cartesian coordinate grid and Scatter coordinate grid.</p>



<p>The following chart shows the Gross Domestic Product and Gross National Income (per Capita) of various countries. The size of the bubble shows the actual GDP of the place.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/47422-bubble.png" alt="" class="wp-image-27017"/></figure></div>



<p></p>

How to Read Scatter Chart and Bubble Chart

<p>A Box Plot is a convenient way of graphically depicting groups of numerical data through their quartiles. They provide a graphical rendition of statistical data based on the minimum, first quartile, median, third quartile, and maximum, also Outliers can be plotted as individual points.</p>



<p>The term "box plot" comes from the fact that the graph looks like a rectangle with lines extending from the top and bottom. Because of the extending lines, this type of graph is sometimes called a box-and-whisker plot. The distances between the different box parts represent the degree of a data dispersion and a data asymmetry to identify outliers. The points values can be compared between themselves (single-series chart) or values inside the category (multi-series chart). In case of several series points are grouped by categories.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/27440-box-plot.png" alt="Box Plot" class="wp-image-27025"/></figure></div>



<h3>Interpreting a Box Plot</h3>



<p>The box plot is interpretted as follows:</p>



<ul><li>The box itself contains the middle 50% of the data. The upper edge (hinge) of the box indicates the 75th percentile of the data set, and the lower hinge indicates the 25th percentile. The range of the middle two quartiles is known as the inter-quartile range.</li><li>The line in the box indicates the median value of the data.</li><li>If the median line within the box is not equidistant from the hinges, then the data is skewed.</li><li>The ends of the vertical lines or "whiskers" indicate the minimum and maximum data values, unless outliers are present in which case the whiskers extend to a maximum of 1.5 times the inter-quartile range.</li><li>The points outside the ends of the whiskers are outliers or suspected outliers.</li></ul>

What is a Box Plot and How to Read It

<h3>Japanese Candlestick Chart</h3>



<p>Japanese Candlestick is a type of financial charts, the modification of OHLC Chart. Candlesticks are usually composed of the body (black or white), and an upper and a lower shadow (wick): the area between the open and the close is called the real body, price excursions above and below the real body are called shadows. The wick illustrates the highest and lowest traded prices of a security during the time interval represented. The body illustrates the opening and closing trades.</p>



<p>If the security closed higher than it opened, the body is white or unfilled, with the opening price at the bottom of the body and the closing price at the top. If the security closed lower than it opened, the body is black, with the opening price at the top and the closing price at the bottom. A candlestick need not have either a body or a wick. The same as for Line Chart, timeline scale is the most common case for the X-axis of the Japanese Candlestick Chart.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/27445-candle-stick.png" alt="Japanese Candlestick Chart" class="wp-image-27028"/></figure></div>



<h3>OHLC Chart</h3>



<p>The OHLC (Open-High-Low-Close) Chart is typically used to illustrate the movement in prices over time. Each data point is depicted as a vertical line representing the price range between the lowest and highest prices over a certain time period, with tick marks on the left and on the right indicating respectively the opening (first) and closing (last) price for the time unit in question. The color of each data point reflects whether the price decreased or increased during that period.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/27449-ohlc.png" alt="" class="wp-image-27029"/></figure></div>



<p></p>

Understanding Japanese Candlestick Charts and OHLC Charts

<h3>Treemap</h3>



<p>Treemaps display hierarchical data by using nested rectangles, that is, smaller rectangles within a larger rectangle.</p>



<p>The Treemap Chart is intended for the visualization of hierarchical data in the form of nested rectangles. Each level of such a tree structure is depicted as a colored rectangle, often called a branch, which contains other rectangles (leaves). The space inside each of the rectangles composing treemaps is highlighted based on the quantitative value of the corresponding data point.</p>



<p>The main advantages of treemap charts conclude in allowing us to do the following:</p>



<ul><li>identify the relationship between two elements of a hierarchical data structure;</li><li>optimize the use of space;</li><li>accurately display multiple elements simultaneously;</li><li>show ratios of each part to the whole;</li><li>visualize attributes by size and color coding.</li></ul>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/27451-treemap.png" alt="Treemap Chart" class="wp-image-27032"/></figure></div>



<p>In this image you have several product groups, such as Food, Vegetables, Dairy and Meet. Each product group consists of a large rectangle. You can regard the product groups as branches of the tree. Within the product groups, you will have product types or products themselves. In this case we have products such as Apples, Bananas, etc.</p>



<p>Today treemaps are a standard tool for visualizing financial data. The screen is divided into rectangular tiles that represent publicly traded companies. The area of a rectangle corresponds to the market capitalization of the company, and the color tells you how the stock price has changed since the previous market close.</p>



<h4>When to use it</h4>



<p>Use a treemap when space is constrained and you have a large amount of hierarchical data that you need to get an overview of. Treemaps should primarily be used with values that can be aggregated.</p>



<h4>Advantages</h4>



<p>Treemaps are economical in that they can be used within a limited space and yet display a large number of items simultaneously.</p>



<p>When there is a correlation between color and size in the tree structure, you are able to see patterns that would be difficult to spot in other ways, for example, when a certain color is particularly relevant.</p>



<h4>Disadvantages</h4>



<p>Treemaps are not good when there is a big difference in the magnitude of the measure values. Nor is a treemap the right choice when mixing absolute and relative values.</p>



<p>Negative values cannot be displayed in treemaps.</p>



<h3>Heatmap</h3>



<p>The Heat Map is a two-dimensional representation of data, in which values are encoded in colors and thus provide viewers with convenient, visualized information. Essentially, this type of chart is a data table with rows and lines denoted by different sets of categories. Each table cell can contain a numerical or logical value that determines the cell color based on a given color palette.</p>



<p>Due to all these peculiarities, heat maps are a convenient data visualization tool for comparing categories, using color to emphasize relationships between data values that would be much harder to understand in a simple table with numbers.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/274546-heatmap.png" alt="Heapmap Chart" class="wp-image-27033"/></figure></div>



<h3>Choropleth Maps</h3>



<p>Choropleth Maps display divided geographical areas or regions that are coloured, shaded or patterned in relation to a data variable. This provides a way to visualise values over a geographical area, which can show variation across the displayed location. When mapping quantitative data, a specific color progression should be used to depict the data properly.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/27456-map1.png" alt="Choropleth Maps" class="wp-image-27034"/></figure></div>



<h3>Bubble Map</h3>



<p>Bubble Maps (or proportional symbol maps) are a class of maps that use the visual variable of size to represent differences in the magnitude of a discrete, abruptly changing phenomenon, e.g. counts of people. Bubble Maps are very similar to Bubble Chart, but instead of a coordinate grid a geographic map is used.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/27456-bubble-map2.png" alt="Bubble Map" class="wp-image-27035"/></figure></div>



<p></p>

Understanding Treemap, Heatmap and Other Map Charts

<p>The advent of data science has opened doors to making discoveries from data in new and innovative ways. By combining statistics, computer science, applied mathematics and data visualizations, a data scientist or analyst can turn vast amounts of complex data into meaningful insights and new knowledge.</p>



<p>The fundamental purpose of any data science project is to use automated methods to analyze massive amounts of data and to extract knowledge from it. This puts data visualization at the forefront of any data science project. We can divide a data science problem into three sub-problems:</p>



<ol><li>Formulate the question</li><li>Understand the answer</li><li>Provide an explanation for the answer</li></ol>



<p>Finding the right question is often considered more important than finding the right answer. This is the reason why the focus of data science should be on both exploratory and explanatory data analysis</p>



<p><strong>Exploratory data analysis</strong>&nbsp;is the critical first step in analyzing the data from an experiment view point. This does not involve any statistical modeling or inference. Instead you just explore the data with an open mind to know more about it. In the process you will explore the data, find mistakes, validate assumptions, determine relationship between variables, and probably come up with important questions that you want to find answers to from the data.</p>



<p><strong>Explanatory data analysis</strong>&nbsp;is where you want to answer specific questions and communicate those answers to the audience. This is where you will apply statistical and mathematical models to infer insights from the data and/or prove a hypothesis. Data visualization plays an important role in both exploratory and explanatory data analysis.</p>



<h3>Why use visualization?</h3>



<p>Data visualization is the most impactful way in which data scientists communicate their findings to their audience. Using data visualizations, you can visually display the data patterns, trends, and relationships that are easy to consume. There are several reasons why visualizations are better than reading numbers or tables.</p>



<ol><li>First, vision is the most powerful communication channel humans possess. We can detect information faster than our eye can move.</li><li>Second, we humans are not very good at detecting patterns from numbers. For example, consider a data set containing the closing prices of five stocks over the past 30 days. It is much easier to identify the trends in these prices if presented as a line chart rather than as a table.</li><li>Third, summary statistics can hide important information.</li><li>Fourth, when information is complex in nature, the right representation can make it simple to understand.</li></ol>



<p>However, we must remember that not all data visualizations are effective and we should spend considerable time in choosing the right visualizations for our data.</p>

Visualization in Data Science

<p>If you want to plot a graph in R for your own consumption, it is pretty easy and you can generally do it with a few lines of code. However, if you are creating the graph for others consumption and want full control over all aspects of the graph, it will require some effort. The standard R distribution comes prepackages with powerful graphing capabilities. There are three in-built graphic packages, namely, Base, Grid and Lattice graphic packages. Apart from these three, there is also ggplot2, another popular graphic package for R.</p>



<h3>Base Graphics</h3>



<p>Base graphics is the oldest and original graphic system built for R. This is the core plotting system and contains functions including&nbsp;<code>plot()</code>,&nbsp;<code>hist()</code>,&nbsp;<code>boxplot()</code>&nbsp;and many others. The code for base graphics is in the graphics package, which is loaded by default when you start R.</p>



<h3>Grid</h3>



<p>Grid is an alternative graphic system which was added to R at a later stage. It implements a completely different graphic system independent of the core graphics package. It is more like a framework and doesn't produce complete graphics by itself. For this reason, we seldom call functions from the grid package directly.</p>



<h3>Lattice</h3>



<p>Lattice is an add-on graphic package that contains code for producing Trellis graphics - graphs that display a variable or the relationship between variables, conditioned on one or more other variables. It includes functions like&nbsp;<code>xyplot()</code>,&nbsp;<code>bwplot()</code>,&nbsp;<code>levelplot()</code>,&nbsp;<code>histogram()</code>. Lattice graphics in R make use of grid graphics. This is incompatible with Base graphics package.</p>



<h3>ggplot2</h3>



<p>ggplot2 is a system for 'declaratively' creating graphics, based on "The Grammar of Graphics". You provide the data, tell 'ggplot2' how to map variables to aesthetics, what graphical primitives to use, and it takes care of the details. Here is the statement of purpose for the ggplo2 package -&nbsp;<em>ggplot2 is a plotting system for R, based on the grammar of graphics, which tries to take the good parts of base and lattice graphics and none of the bad parts. It takes care of many of the fiddly details that make plotting a hassle (like drawing legends) as well as providing a powerful model of graphics that makes it easy to produce complex multi-layered graphics.</em></p>



<p>Base graphics are used most commonly and are a very powerful system for creating 2-D graphics. In this course, we will focus on the Base graphics system and ggplot2.</p>

Graphic Systems in R

<p>R comes with many built-in datasets which are quite useful while learning R. To begin learning the basics of data visualization in R, we will make use of some of these datasets.</p>



<h3>Datasets in the 'datasets' package</h3>



<p>Many datasets are included in a package called&nbsp;<code>datasets</code>&nbsp;which is distributed with R so these datasets are instantly available to you for use. For example, two datasets namely&nbsp;<code>cars</code>&nbsp;and&nbsp;<code>pressure</code>&nbsp;are included in this default datasets package. So, you can access their data by using functions such as&nbsp;<code>head(cars)</code>,&nbsp;<code>summary(cars)</code>, etc. The following examples show results of calls to these functions:</p>



<pre class="wp-block-code"><code>> head(cars)
  speed dist
1     4    2
2     4   10
3     7    4
4     7   22
5     8   16
6     9   10
></code></pre>



<pre class="wp-block-code"><code>> summary(cars)
     speed           dist       
 Min.   : 4.0   Min.   :  2.00  
 1st Qu.:12.0   1st Qu.: 26.00  
 Median :15.0   Median : 36.00  
 Mean   :15.4   Mean   : 42.98  
 3rd Qu.:19.0   3rd Qu.: 56.00  
 Max.   :25.0   Max.   :120.00  
></code></pre>



<pre class="wp-block-code"><code>> head(pressure)
  temperature pressure
1           0   0.0002
2          20   0.0012
3          40   0.0060
4          60   0.0300
5          80   0.0900
6         100   0.2700
></code></pre>



<pre class="wp-block-code"><code>> summary(pressure)
  temperature     pressure       
 Min.   :  0   Min.   :  0.0002  
 1st Qu.: 90   1st Qu.:  0.1800  
 Median :180   Median :  8.8000  
 Mean   :180   Mean   :124.3367  
 3rd Qu.:270   3rd Qu.:126.5000  
 Max.   :360   Max.   :806.0000  
></code></pre>



<p>To learn more about a dataset, you can use the help function as&nbsp;<code>help(cars)</code>.</p>



<p>If you want to get a list of all the datasets, you can do so using the&nbsp;<code>data()</code>&nbsp;function.</p>



<h3>Datasets in Other Packages</h3>



<p>Any R package can choose to include datasets. You can access the data from a package using the<code>data()</code>&nbsp;function by using the package argument as follows:</p>



<pre class="wp-block-code"><code>data(datasetname, package="packagename")</code></pre>



<p>For example, there's a popular package called <code>MASS</code> which contains datasets (such as Cars93). We can access the Cars93 dataset by calling the <code>data()</code> function.</p>



<pre class="wp-block-code"><code>> data(Cars93, package="MASS")</code></pre>



<p>After this call to <code>data()</code>, the Cars93 dataset is available for use in R.</p>

Accessing Built-in Datasets in R

<p>When you start analyzing a new dataset, your first requirement would be to know the variables in the dataset and the relationship between them. A scatter plot is the perfect place to start with. It is the quickest way to view the relationship between any two variables x and y.</p>



<p>You can create a scatter plot using the generic <code>plot()</code> function in R.</p>



<pre class="wp-block-code"><code>plot(x,y)</code></pre>



<p>The function itself doesn't return anything back to the console but instead draws the plot in the plot window.</p>



<p>The two variables, x and y, could be two separate vectors or it could be a dataframe with two columns.</p>



<p>The following example shows the scatter plot created using the <code>cars</code> dataset. The data gives the speed of cars and the distances taken to stop. It has two columns, speed and dist. The first column, speed, becomes the x-axis and the second column, dist, becomes the y-axis.</p>



<pre class="wp-block-code"><code>plot(cars)</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/28720-scatter-plot.png" alt="" class="wp-image-27074"/></figure></div>



<p>As you can see, the graph draws a point for each pair of speed and dist. The general observation from the scatter plot is that the higher the speed, the higher is the distance to stop.</p>



<p>If the dataset contains more than 2 columns, the&nbsp;<code>plot()</code>&nbsp;function will return multiple scatter plots each representing relationship between two variables.</p>



<p>Let's take another dataset called&nbsp;<code>whiteside</code>&nbsp;from the&nbsp;<code>MASS</code>&nbsp;package. The dataset contains home insulation data containing three variables, namely, Insul, Temp and Gas.</p>



<ul><li><strong>Insul:</strong>&nbsp;A factor, before or after insulation.</li><li><strong>Temp:</strong>&nbsp;Purportedly the average outside temperature in degrees Celsius.</li><li><strong>Gas:</strong>&nbsp;The weekly gas consumption in 1000s of cubic feet.</li></ul>



<p>If we call the <code>plot()</code> function on this dataset, it will plot multiple scatter plots representing relationship between all three variables.</p>



<pre class="wp-block-code"><code>> #load the data
> data(whiteside,package="MASS")
> head(whiteside)
   Insul Temp Gas
1 Before -0.8 7.2
2 Before -0.7 6.9
3 Before  0.4 6.4
4 Before  2.5 6.0
5 Before  2.9 5.8
6 Before  3.2 5.8
> #plot the data
> plot(whiteside)</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/28720-multi-scatter-plot.png" alt="Scatter Plot in R" class="wp-image-27075"/></figure></div>



<p>Each scatter plot draws points for two variables. For example, the graph in the lower left corner has Insul on x-axis and Gas on y-axis. Similarly, the graph in the bottom row, middle column has Temp on x-axis and Gas on y-axis.</p>



<p>The&nbsp;<code>plot()</code>&nbsp;function we use above is generic function, that is, it will change its behavior depending on the types of arguments provided and will produce different results. We just saw the use of the&nbsp;<code>plot()</code>&nbsp;function in its most basic form. In the following lessons, we will see how we can customize the plots and enhance them using various arguments.</p>



<h3>Exercise</h3>



<p>Load the <code>Cars93</code> dataset from the <code>MASS</code> package and use the <code>plot()</code> function to draw scatter plots on its variables.</p>



<h2>Enhancing a Plot</h2>



<p>Now that we know how to create a basic plot using the <code>plot()</code> function, let's learn how we can enhance the chart in various ways. We will start with a new dataset which contains daily stock returns for five stocks, namely, Goldman Sachs, Citi, Apple, Facebook and JC Penny for a period of one year. The data is provided in csv format, so we will first load it into R.</p>



<div class="wp-block-file"><a href="https://financetrain.com/wp-content/uploads/stock_returns.csv">stock_returns</a><a href="https://financetrain.com/wp-content/uploads/stock_returns.csv" class="wp-block-file__button" download>Download</a></div>



<h3>Load the Data</h3>



<p>I've placed the data file in my working directory and then used the <code>read.csv()</code> function to load the data into an R dataframe called 'stock_returns'.</p>



<pre class="wp-block-code"><code>> getwd()
&#91;1] "C:/Users/Manish/Documents"
> setwd("C:/r-programming/data")
> getwd()
&#91;1] "C:/r-programming/data"
> stock_returns &lt;- read.csv("stock_returns.csv")
> head(stock_returns)
        Date    gs    c  aapl    fb   jcp
1 16/04/2015 -0.44 1.52 -0.48 -0.47 -2.69
2 15/04/2015  1.71 0.91  0.38 -0.98 -2.40
3 14/04/2015  1.09 0.13 -0.43  0.61 -2.66
4 13/04/2015 -0.03 0.44 -0.20  1.18  1.95
5 10/04/2015  0.38 0.58  0.43 -0.16  0.22
6 09/04/2015  1.21 0.46  0.76 -0.13  1.32</code></pre>



<h3>Create a Scatter Plot</h3>



<p>If we just call the <code>plot()</code> function on 'stock_returns' dataset, it will plot multiple scatter plots one for each pair of columns. However, for our use, we will just create a scatter plot for GoldmanSachs and Citi's stock returns. We can do so by supplying x and y values separately as shown below:</p>



<pre class="wp-block-code"><code>> plot(stock_returns$gs,stock_returns$c)</code></pre>



<p>The resulting plot is shown below:</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/28725_stock_returns_scatter.png" alt="Scatter Plot in R" class="wp-image-27077"/></figure></div>



<h3>Add Title and Axis Labels</h3>



<p>The scatter plot we created is quite plain. We can make it more readable by adding a title and labels to the X and Y axis. We will use the&nbsp;<code>plot()</code>&nbsp;function arguments to do so.</p>



<ul><li>The&nbsp;<code>main</code>&nbsp;argument for the title</li><li>The&nbsp;<code>xlab</code>&nbsp;argument for the x-axis</li><li>The&nbsp;<code>ylab</code>&nbsp;argument for the y-axis</li></ul>



<p>After adding the title and labels, we can also add grid to the graph by calling the <code>grid()</code> function after calling the <code>plot()</code> function.</p>



<pre class="wp-block-code"><code>> plot(stock_returns$gs,stock_returns$c, main="Scatter Plot: 1-year Daily Returns", xlab="GoldmanSachs Returns", ylab="Citigroup Returns")
> grid()</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/28725-scatter-2.png" alt="Daily Returns Scatter Chart" class="wp-image-27078"/></figure></div>



<p>In the above example, we first plotted the graph and then added the grid to it. The alternative (and preferred) method is to plot the graph using the&nbsp;<code>plot()</code>&nbsp;function but with the argument&nbsp;<code>type=n</code>&nbsp;which will prevent the graph from printing. Then we call the&nbsp;<code>grid()</code>&nbsp;function to add the grid, and then finally call the low-level graphics function such as&nbsp;<code>points()</code>&nbsp;or&nbsp;<code>lines()</code>&nbsp;to overlay the graph on the grid.</p>



<h3>Plot a Regression Line</h3>



<p>We can add a regression line to this scatter plot of returns for GoldmanSachs and Citigroup as shown below:</p>



<p>1. Perform a linear regression using lm() on the two variables. lm stands for "linear model"</p>



<pre class="wp-block-code"><code>m &lt;- lm(stock_returns$c ~ stock_returns$gs)</code></pre>



<p>2. Draw the scatter plot</p>



<pre class="wp-block-code"><code>plot(stock_returns$c ~ stock_returns$gs, main="Scatter Plot: 1-year Daily Returns", xlab="GoldmanSachs Returns", ylab="Citigroup Returns")</code></pre>



<p>3. Add the regression line using <code>abline</code> function.</p>



<pre class="wp-block-code"><code>abline(m)</code></pre>



<p>The graph will now look as follows:</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/28725-scatter-regression.png" alt="Scatter Chart with Regression Line" class="wp-image-27079"/></figure></div>



<p>Notice that we defined the plot as&nbsp;<code>plot(stock_returns$c ~ stock_returns$gs)</code>. This is to keep the order of variables similar to how it is in the&nbsp;<code>lm()</code>&nbsp;function. The alternative way is&nbsp;<code>plot(stock_returns$gs,stock_returns$c)</code>&nbsp;that we used earlier.</p>



<h3>Exercise</h3>



<p>Load the 'stock_returns' dataset into R and create a scatter plot with Apple's returns on x-axis and Facebook's returns on y-axis. Then add a title, axis labels and a regression line to the plot.</p>

How to Create a Scatter Plot in R

<p>Let's say you have Sales Orders data for a sports equipment manufacturer and you want to plot the Revenue and Gross Margins on a scatter plot. However, you also have a ProductLine column that contains information about the product category and you want to distinguish the x,y points by the ProductLine.</p>



<p>We can do so using the <code>pch</code> argument of the <code>plot</code> function.</p>



<pre class="wp-block-code"><code>> plot(x, y, pch=as.integer(f))</code></pre>



<p>By specifying this option, the plot will use a different plotting symbol for each point based on its group (f).</p>



<p>We have created a sample dataset for this lesson which contains Sales, Gross Margin, ProductLine and some more factor columns. You can download this dataset from the Lesson Resources section. As always, we will first load the dataset into an R dataframe.</p>



<div class="wp-block-file"><a href="https://financetrain.com/wp-content/uploads/Sales_Products.csv">Sales_Products</a><a href="https://financetrain.com/wp-content/uploads/Sales_Products.csv" class="wp-block-file__button" download>Download</a></div>



<pre class="wp-block-code"><code>> setwd("C:/r-programming/data")
> getwd()
> sales&lt;-read.csv("Sales_Products.csv")</code></pre>



<p>Before plotting the graph, it's a good idea to learn more about the data by using the&nbsp;<code>summary()</code>&nbsp;and&nbsp;<code>head()</code>&nbsp;functions.</p>



<p>We are interested in three columns from this dataset:</p>



<ul><li><strong>Revenue:</strong>&nbsp;The total revenue from each order. We will plot this on x-axis.</li><li><strong>Gross Margin:</strong>&nbsp;The gross margin from each order. We will plot this on y-axis.</li><li><strong>ProductLine:</strong>&nbsp;The product category. We will group the data by ProductLine.</li></ul>



<p>We can now draw the scatter plot using the following command:</p>



<pre class="wp-block-code"><code>plot(sales$Revenue,sales$GrossMargin,pch=as.integer(sales$ProductLine))</code></pre>



<p>The result is displayed below. You can clearly see the points with different symbols according to their group.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/28735-scatter-group-1.png" alt="" class="wp-image-27137"/></figure></div>



<p>Notice that R has converted the y-axis scale values to scientific notation. We can correct this by changing the option <code>scipen </code>to a higher value. This controls which numbers are printed in scientific notation.</p>



<pre class="wp-block-code"><code>> options("scipen" = 10)
> options()$scipen
&#91;1] 10</code></pre>



<p>If you plot the chart again, the numbers would display correctly.</p>



<figure class="wp-block-image size-large"><img src="https://financetrain.com/wp-content/uploads/28735-scatter-2.png" alt="" class="wp-image-27138"/></figure>



<h3>Add a Legend</h3>



<p>Now that we have different symbols being used for different groups, we can make the graph even more convenient by adding a legend to it. We can do so by calling the <code>legend</code> function after the plot function.</p>



<pre class="wp-block-code"><code>legend(x, y=NULL, legend, …)</code></pre>



<p>x, y are the coordinates for the legend box. There are two ways to specify x: 1) Specify the position by using "topleft", "topright", etc. 2) Use an x-coordinate for the top-left corner of the legend. If you choose option 1 for specifying x, then y can be skipped. Alternatively you need to specify the y-coordinate for the top-left corner of the legend.</p>



<p>The third argument "legend" is a vector of the character strings to appear in the legend.</p>



<p>You also need to specify a fourth argument that varies depending on what you're labeling. You can create legends for points, lines, and colors. In our case, we are creating legend for points, so we will provide the forth argument <code>pch</code> which is also a vector indicating that we are labeling the points by their type.</p>



<pre class="wp-block-code"><code>> legend("topleft", c("Camping Equipment","Golf Equipment","Mountaineering Equipment", "Outdoor Protection", "Personal Accessories"), pch=1:5)</code></pre>



<p>The graph will now look as follows:</p>



<figure class="wp-block-image size-large"><img src="https://financetrain.com/wp-content/uploads/28735-scatter-legend.png" alt="" class="wp-image-27139"/></figure>



<p>The legend function can also create legends for colors, fills, and line widths.The <code>legend()</code> function takes many arguments and you can learn more about it using help by typing <code>?legend</code>.</p>



<h3>Exercise</h3>



<ul><li>Download and load the&nbsp;<code>Sales_Products</code>&nbsp;dataset in your R environment</li><li>Use the&nbsp;<code>summary()</code>&nbsp;function to explore the data</li><li>Create a scatter plot for Sales and Gross Margin and group the points by&nbsp;<code>OrderMethod</code></li><li>Add a legend to the scatter plot</li><li>Add different colors to the points based on their group. (Hint: Use the&nbsp;<code>col</code>&nbsp;argument in the&nbsp;<code>plot()</code>&nbsp;function</li></ul>

Create a Scatter Plot in R with Multiple Groups

<p>In R, we can create a bar plot using the&nbsp;<code>barplot()</code>&nbsp;function. A Bar Plot or Bar Graph is primarily used to compare values. It presents grouped data using rectangular bars whose lengths are proportional to the values that they represent. Let's take our Product Sales data where we have the Revenue and Gross Margin for each order along with various attributes such as ProductLine, RetailerType, OrderMethod etc. We can use this data to create a bar chart which plots the total or average sales on y-axis as bars and the x-axis as one of these factors that we are interested in such as ProductLine.</p>



<p>The first step is to group the Total Sales by ProductLine using the&nbsp;<code>tapply()</code>&nbsp;function as shown below:</p>



<pre class="wp-block-code language-r"><code>> sales_by_productline &lt;- tapply(sales$Revenue, sales$ProductLine, sum)</code></pre>



<p>This gives us the sales to be plotted as bars. We can now plot the chart using the&nbsp;<code>barplot()</code>&nbsp;function.</p>



<pre class="wp-block-code language-r"><code>> barplot(sales_by_productline)</code></pre>



<p>The above function call will create the bar chart. We can enhance the chart by adding a title and axis labels.</p>



<pre class="wp-block-code language-r"><code>> barplot(sales_by_productline, main="Sales by ProductLine", xlab="ProductLine",ylab="$")</code></pre>



<p>The resulting bar plot is displayed below:</p>



<figure class="wp-block-image size-large"><img src="https://financetrain.com/wp-content/uploads/28745-bar-chart.png" alt="" class="wp-image-27144"/></figure>



<h3>Adding Colors to Bars</h3>



<p>We can add different colors to the bars in the bar plot by adding the&nbsp;<code>col</code>&nbsp;argument. The&nbsp;<code>col</code>&nbsp;argument is basically a vector of colors. R has some inbuilt functions to generate vectors of colors, for example, the&nbsp;<code>gray</code>&nbsp;function generates a vector of grays. Similarly the&nbsp;<code>rainbow</code>&nbsp;function generates a vector of rainbow colors. Alternatively you can explicitly supply a vector containing color codes. In the following example, we extend our bar plot by painting the bars with rainbow colors:</p>



<pre class="wp-block-code language-r"><code>> barplot(sales_by_productline, col=rainbow(5), main="Sales by ProductLine", xlab="ProductLine",ylab="$")</code></pre>



<p>The graph will now look as follows:</p>



<figure class="wp-block-image size-large"><img src="https://financetrain.com/wp-content/uploads/28745-color-bar-chart.png" alt="" class="wp-image-27145"/></figure>



<p>To supply explicit colors, you would pass the col argument as&nbsp;<code>col = c("Red", "Green", "Blue", ...)</code>. Alternatively, you can also specify the hex codes for the colors.</p>



<p>The base graphics library provides only basic features for plotting charts. For example, in the above bar chart, we could have plotted mean revenue values instead of totals. That would have worked fine. However, suppose we also wanted to plot max and min sales for each product line as markers above and below the mean bars. That would be very complicated to achieve here. However, the same thing could be easily done with other plotting libraries such as&nbsp;<code>gplots</code>. We will learn about the&nbsp;<code>gplots</code>&nbsp;library in a separate course.</p>



<h3>Exercise</h3>



<ul><li>Load the Product Sales data in your R environment</li><li>Create a bar plot containing the Average Gross Margin, grouped by RetailerType</li><li>Color by bars with various shades of gray according to their rank, shorter bars being lighter colors and taller bars being darker colors.</li></ul>

Creating a Bar Chart in R

<p>In R, we can create a line plot using the same&nbsp;<code>plot()</code>&nbsp;function by adding a plot type of "l". This will plot the (x,y) paired observations and connect them with lines.</p>



<pre class="wp-block-code language-r"><code>plot(x,y,type="l")</code></pre>



<p>Let's generate our own data for this lesson. We will use the&nbsp;<code>rnorm()</code>&nbsp;function to generate a set of 100 random numbers that follow a <a href="https://financetrain.com/standard-normal-distribution/">normal distribution</a>. These random numbers will be plotted on y-axis. x-axis will be a sequence of numbers from 1 to 100.</p>



<pre class="wp-block-code language-r"><code>> x &lt;- c(1:100)
> rand_data &lt;- rnorm(100, mean = 0)
> data &lt;- data.frame(x, rand_data)
> plot(data$x, data$rand_data, type = "l", xlab="x", ylab = "Data")</code></pre>



<p>The resulting line plot is displayed below:</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/28750-line-plot.png" alt="" class="wp-image-27161"/><figcaption>Line Chart</figcaption></figure></div>



<h3>Line Type</h3>



<p>You can change the appearance of the line by using the&nbsp;<code>lty</code>&nbsp;parameter.</p>



<ul><li>lty="solid" or lty=1 (default)</li><li>lty="dashed" or lty=2</li><li>lty="dotted" or lty=3</li><li>lty="dotdash" or lty=4</li><li>lty="longdash" or lty=5</li><li>lty="twodash" or lty=6</li><li>lty="blank" or lty=0</li></ul>



<h3>Line Width</h3>



<p>The line width can be changed using the&nbsp;<code>lwd</code>&nbsp;parameter. The default width is 1. So, you can plot a thicker line using a higher number.</p>



<h3>Line Color</h3>



<p>The line color can be changed using the&nbsp;<code>col</code>&nbsp;parameter.</p>



<p>Below we replot the line plot with a dotted line, thickness of 2, and in orange color.</p>



<pre class="wp-block-code language-r"><code>> plot(data$x, data$rand_data, type = "l", xlab="x", ylab = "Data",lty=3,lwd=2,col="orange")</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/28750-line-plot-dotted.png" alt="" class="wp-image-27162"/><figcaption>Line Chart</figcaption></figure></div>



<h3>Exercise</h3>



<p>We know that numerical data generally conforms to a normal <a href="https://financetrain.com/probability-distribution/">probability distribution</a> characterized by a bell curve. In this exercise you are asked to create a bell curve using the normal data.</p>



<ul><li>Use the&nbsp;<code>rnorm()</code>&nbsp;function to generate random numbers that follow a normal distribution (mean 0 and standard deviation of 1). Generate upto 5000 numbers. Store these numbers in a variable x.</li><li>Plot the density of these random numbers using&nbsp;<code>plot(density(x))</code>&nbsp;This will plot the bell curve.</li><li>Try different values for mean and standard deviation and observe how the graph changes.</li></ul>



<p><em>Note: You can also compute the Gaussian density using the&nbsp;<code>dnorm()</code>&nbsp;function, for example&nbsp;<code>dnorm(x, mean = 0, sd = 1)</code>.</em></p>

Creating a Line Chart in R

<p>It's a common scenario to plot multiple datasets together on a single graph. For example, we may want to plot the daily returns from multiple stocks on a single chart to understand how they trend vis-a-vis each other. Similarly, we may want to plot multiple normal distribution curves with different mean and standard deviations.</p>



<p>To plot multiple datasets, we first draw a graph with a single dataset using the&nbsp;<code>plot()</code>&nbsp;function. Then we add the second data set using the&nbsp;<code>points()</code>&nbsp;or&nbsp;<code>lines()</code>&nbsp;function.</p>



<p>Let's learn this with the help of an example where we will plot multiple normal distribution curves.</p>



<h3><strong>Generate x-axis data</strong></h3>



<p>First we will generate data for x-axis which will be a sequence of 200 evenly spaced numbers ranging from -5 to 5. We can do this using the&nbsp;<code>seq()</code>&nbsp;function in R.</p>



<pre class="wp-block-code language-r"><code>> x&lt;-seq(-5,5,length=200)</code></pre>



<h3><strong>Calculate Values for Normal Distribution</strong></h3>



<p>We can do this in two ways: 1) Generate random numbers using&nbsp;<code>rnorm()</code>&nbsp;and then apply the&nbsp;<code>density()</code>&nbsp;function to the data. 2) Alternatively we can do this directly using the&nbsp;<code>dnorm()</code>&nbsp;function which gives the density of the distribution function. In more simple terms, this function gives height of the probability distribution at each point for a given mean and standard deviation. For our purpose, we will generate multiple datasets with different means and standard deviations.</p>



<pre class="wp-block-code language-r"><code>> y1&lt;-dnorm(x,mean=0,sd=0.2)
> y2=dnorm(x,mean=2,sd=0.5)
> y3&lt;-dnorm(x,mean=-2,sd=0.8)</code></pre>



<h3><strong>Combine Datasets</strong></h3>



<p>We can also combine all the data into a single dataframe (optional).</p>



<pre class="wp-block-code language-r"><code>> data&lt;-data.frame(x,y1,y2,y3)</code></pre>



<h3><strong>Plot the First Curve</strong></h3>



<pre class="wp-block-code language-r"><code>> plot(data$x,data$y1,type="l",main="Normal Distribution",xlab="x",ylab="y")</code></pre>



<p>The plot looks as follows:</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/28756-curve-1-1.png" alt="" class="wp-image-27169"/><figcaption>Normal Distribution</figcaption></figure></div>



<h3><strong>Add Lines for the Second Normal Density</strong></h3>



<p>We can now add the lines for the second and third density using the&nbsp;<code>lines()</code>&nbsp;function.</p>



<pre class="wp-block-code language-r"><code>> lines(data$x,data$y2,lty=2,lwd=2,col="green")
> lines(data$x,data$y3,lty=3,lwd=2,col="blue")</code></pre>



<p>The resulting graph is displayed below:</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/28756-curve-2.png" alt="" class="wp-image-27170"/></figure></div>



<p>Note that if we were plotting just the scatter graph without lines, we could add more data points to it using the&nbsp;<code>points()</code>&nbsp;function instead of the&nbsp;<code>lines()</code>&nbsp;function.</p>



<h3>Setting Canvas Size</h3>



<p>Sometimes when we want to add multiple datasets to a single plot, it is important to correctly specify the size of the canvas. Let's say the first dataset that you plot has an x-value range of 0 to 100. Once this is plotted, the graph will draw the x-axis with the 0-100 range. However, assume now that the second dataset that you want to plot has x values ranging from 0 to 200. Since the initial plot doesn't consider this, the points from the second dataset will be plotted off the chart and will be cut-off. To correct this problem, we need to set the coordinates for the graph in the beginning itself. This can be done using the xlim and ylim arguments.</p>



<p>Suppose we want to plot two datasets (x1,y1) and (x2,y2). We can compute the limits using the range function and then set them using xlim and ylim.</p>



<pre class="wp-block-code language-r"><code>> xlim &lt;- range(c(x1,x2))
> ylim &lt;- range(c(y1,y2))
> plot(x1, y1, type="l", xlim=xlim, ylim=ylim)
> lines(x2, y2, lty="dashed")</code></pre>

Plotting Multiple Datasets on One Chart in R

<p>We've already seen how we can enhance the plots created in R by adding title, axis labels, legends, etc. Let's learn about a few more enhancements.</p>



<h3>Adding Vertical/Horizontal Lines to the Graph</h3>



<p>You may want to add a vertical or a horizontal line to your graph at a certain axis point. This can be done using the&nbsp;<code>abline()</code>&nbsp;function.</p>



<pre class="wp-block-code language-r"><code>> abline(v=0) # Draw a vertical line at x = 0.
> abline(h=0) # Draw a vertical line at y = 0.</code></pre>



<p>In the below example, we add a vertical line at x=0 on our normal distribution plot.</p>



<pre class="wp-block-code language-r"><code>> abline(v=0,col="red",lwd=2)</code></pre>



<p>The resulting graph is displayed below:</p>



<figure class="wp-block-image size-large"><img src="https://financetrain.com/wp-content/uploads/28762-vline.png" alt="" class="wp-image-27174"/><figcaption>Normal Distribution - Vertical Line</figcaption></figure>



<h3>Adding Explanatory text to the Visualization</h3>



<p>To make a data visualization useful for its users, it is important to add sufficient explanatory text to it. The function&nbsp;<code>text()</code>&nbsp;allows us to add text anywhere inside the plotting area.</p>



<p>The below code adds the text "Normal Distribution Curves" to our distributions graph.</p>



<pre class="wp-block-code language-r"><code>> text(-3, 1.5, "Normal Distribution Curves")</code></pre>



<figure class="wp-block-image size-large"><img src="https://financetrain.com/wp-content/uploads/28762-text.png" alt="" class="wp-image-27175"/><figcaption>Normal Distribution - Adding Text</figcaption></figure>



<p>The&nbsp;<code>text()</code>&nbsp;function can be used to add text in many ways. Use the help&nbsp;<code>?text</code>&nbsp;to learn more about the function.</p>



<h3>Using Colors Effectively</h3>



<p>While designing data visualizations, it is important to use colors to easily distinguish between various data points. For example, if you have two lines on a plot, it's not a good idea to draw one as dark green and another as dark blue. It will be difficult to distinguish between the two lines. Iliinsky and Steele recommend a set of 12 colors, in descending order of desirability from the top of the plot to the bottom.</p>



<pre class="wp-block-code language-r"><code>> # Iliinsky and Steele color name vector
> colors &lt;- c("red", "green", "yellow", "blue", "black", "white", "pink", "cyan", "gray", "orange", "brown", "purple")
> data &lt;- c(rep(3, 4), rep(2, 4),rep(1,4))
> barplot(rev(data), horiz = TRUE, col = rev(colors),names.arg = rev(colors), las = 1)
</code></pre>



<p>As you can see, we are using <code>rep()</code> in a function where we are creating our own sample data: <code>rep()</code> function replicates the values in x. For example, <code>rep(3, 4)</code> will repeat the value 3 4 times, i.e., 3 3 3 3. <code>rev()</code> function reverses an R object, including vector, array etc.</p>



<figure class="wp-block-image size-large"><img src="https://financetrain.com/wp-content/uploads/28762-colors.png" alt="" class="wp-image-27176"/><figcaption>R Chart- Adding Colors</figcaption></figure>



<p>Note that in the above graph, we use the argument&nbsp;<code>las=1</code>&nbsp;to rotate the plot's axis labels. The&nbsp;<code>las</code>&nbsp;argument can have three possible values based on which the placement of the label will differ. With&nbsp;<code>las</code>&nbsp;equal to 0, the label will always be parallel to the axis (default); With&nbsp;<code>las</code>&nbsp;equal to 1, the label will be put horizontally. Value 2 will make it perpendicular to the axis and 3 will place it vertically.</p>



<p><code>las</code>&nbsp;is actually an argument of the&nbsp;<code>axis()</code>&nbsp;function, that is, you can call the&nbsp;<code>axis()</code>&nbsp;function after the plot function to make changes to the axis. Apart from rotating the axis labels, it is capable of many things, such as changing the axis position, for example&nbsp;<code>axis(3)</code>&nbsp;will draw the x-axis above the plot area and&nbsp;<code>axis(4)</code>&nbsp;will draw the y-axis on the right side of the plot area. Note that if you want to edit the axis settings after plotting the graph using&nbsp;<code>axis()</code>&nbsp;function, it's a good idea to set the default axis to false (axes=FALSE) during the initial plot creation.</p>

Adding Details and Features to R Plots

<p>We have already learned about how we can create meaningful data visualizations in R using the Base Graphics package. In this section, we will learn about how we can create even more powerful data graphics using an R package called&nbsp;<code>ggplot2</code>.</p>



<h3>What is ggplot2?</h3>



<p>The ggplot2 package is a very popular alternative to the base graphics package in R with over 1 million downloads in a year. At its core, it is an R package for creating statistical (data) graphics. However, it is different from other data visualization packages because it implements a very strong underlying grammar for creating these graphics. The ggplot2 package is an implementation of the ideas in the book, The Grammar of Graphics, by Leland Wilkinson, whose goal was to set out a set of general unifying principles for the visualization of data.</p>



<p>The package has very few base functions that makes it easy to learn and use. However, based on the grammar of graphics, we can combine these functions in many different ways to produce many different types of graphics.&nbsp;<code>ggplot2</code>&nbsp;is also very good at setting reasonable default values enabling users to create good looking, hassle-free graphs. For example, it automatically adds legends to the graphs. Defaults enable us to use ggplot2 without knowing the grammar. However, knowing grammar allows us to build graphs from concepts rather than recall of commands and options. It also enables us to create new and improved graphs.</p>



<p>One of the important ideas in&nbsp;<code>ggplot2</code>&nbsp;is that it allows us to build the graphic iteratively, one layer at a time. We can start with one layer that plots the raw data. Then add more layers showing annotations or statistical summaries. This matches with how we analyze data and think about data visualizations making it easy for use to create complex graphics iteratively.</p>



<h3>Installation and Use</h3>



<p>The first thing you need to do is install the <code>ggplot2</code> package. This is done using the <code>install.packages()</code> function as shown below:</p>



<pre class="wp-block-code"><code>> install.packages("ggplot2")</code></pre>



<p>This command will install the&nbsp;<code>ggplot2</code>&nbsp;package in your R instance along with any dependency packages</p>



<p>If you're using RStudio, you will see the newly installed package listed under packages in the bottom right window.</p>



<figure class="wp-block-image size-large"><img src="https://financetrain.com/wp-content/uploads/ggolot2-1.png" alt="" class="wp-image-27197"/></figure>



<p>Once installed you can load it in your current R session using the following command:</p>



<pre class="wp-block-code"><code>> library("ggplot2")</code></pre>



<h3>ggplot2 Documentation</h3>



<ul><li><a href="https://ggplot2.tidyverse.org/reference/" target="_blank" label="undefined (opens in a new tab)" rel="noreferrer noopener nofollow">https://ggplot2.tidyverse.org/reference/</a></li><li>Contains help files for most (all?) ggplot2 functions</li><li>Help files typically contain numerous code and graphics examples</li></ul>



<h3>Plotting with ggplot2</h3>



<p><code>ggplot2</code>&nbsp;offers two ways to to produce plot objects: 1)&nbsp;<code>qplot()</code>&nbsp;and 2)&nbsp;<code>ggplot()</code></p>



<p>The&nbsp;<code>qplot()</code>&nbsp;(for quick plot) can be used to create the most common graph types. It hides much of the complexity when creating standard graphs.</p>



<p>The&nbsp;<code>ggplot()</code>&nbsp;function on the other hand brings the full power of grammar of graphics. It has slightly steeper learning curve but allows much more flexibility when building graphs</p>



<p>Our focus in this module will be on creating visualizations using the&nbsp;<code>ggplot()</code>&nbsp;function.</p>



<h3>Example</h3>



<p>The following example uses the&nbsp;<code>Insurance</code>&nbsp;dataset from the&nbsp;<code>MASS</code>&nbsp;package. Below we show a very basic graph created using the&nbsp;<code>qplot()</code>&nbsp;and&nbsp;<code>ggplot()</code>&nbsp;functions using this data. The data given in data frame&nbsp;<code>Insurance</code>&nbsp;consist of the numbers of policyholders of an insurance company who were exposed to risk, and the numbers of car insurance claims made by those policyholders in the third quarter of 1973. We plot a simple scatter chart with the No. of policy holders on x-axis and No. of claims on y-axis.</p>



<p>Since we're using the most basic settings, both the functions will produce the same chart.</p>



<p><strong>Load the dataset</strong></p>



<pre class="wp-block-code"><code>> library(MASS)
> data(Insurance)</code></pre>



<p><strong>Inspect data using the str() function</strong></p>



<pre class="wp-block-code"><code>> str(Insurance)</code></pre>



<p>Plot the chart using <code>qplot() </code>or <code>ggplot()</code></p>



<p><strong>Format in qplot()</strong></p>



<pre class="wp-block-code"><code>> qplot(Insurance$Holders ,Insurance$Claims)</code></pre>



<p><strong>Format in ggplot()</strong></p>



<pre class="wp-block-code"><code>> ggplot(Insurance, aes(x = Holders, y = Claims))+geom_point()</code></pre>



<p>Both these commands will draw the scatter plot with the Holders on x-axis and Claims on y-axis.</p>



<figure class="wp-block-image size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-intro1.png" alt="" class="wp-image-27198"/></figure>



<p>As you can see&nbsp;<code>ggplot</code>&nbsp;has automatically taken care of most of the details such as axis, legends, etc.</p>



<p>In the following lessons, we will learn about the grammar of graphics and use it to create interesting data visualizations on some financial datasets.</p>

Introduction to ggplot2

<p>We know that the grammar of a language is a set of rules that explain how words are used in a language to form meaningful sentences. Similarly, grammar of graphics defines rules for structuring mathematics and aesthetic elements into a meaningful graph, i.e., graphics are built based on an underlying grammar.</p>



<p>The grammar of graphics in&nbsp;<code>ggplot</code>&nbsp;is a plotting framework developed by Leland Wilkinson in his book 'Grammar of Graphics' published in 1999.</p>



<p>There are two important principles here:</p>



<ol><li>Graphics are made of distinct layers of grammatical elements</li><li>Plots are built with appropriate aesthetic mappings to makes these plots meaningful</li></ol>



<blockquote class="wp-block-quote"><p>In brief, the grammar tells us that a statistical graphic is a mapping from data to aesthetic attributes (colour, shape, size) of geometric objects (points, lines, bars). The plot may also contain statistical transformations of the data and is drawn on a specific coordinate system</p><cite>From ggplot2 book</cite></blockquote>



<h3>Components of Grammar of Graphics</h3>



<p>There are 7 grammatical elements in&nbsp;<code>ggplot2</code>&nbsp;with 3 being essential. The following table summarizes these elements:</p>



<figure class="wp-block-table"><table><tbody><tr><td><strong>Element</strong></td><td><strong>Description</strong></td></tr><tr><td>Data</td><td>This is the dataset being plotted containing the variables to be plotted on the graph.<em>This is an essential element.</em></td></tr><tr><td>Aesthetics</td><td>Aesthetics refers to the scales on which we map our data. For example, we will map one variable to the x-axis, and another to y-axis. We may use to differentiate different attributes by color. Some common aesthetics to consider are axis, shape, size and color.<em>This is an essential element.</em></td></tr><tr><td>Geometries</td><td>Geom refers to the actual visual elements used for the data in the plot, such as points, lines, and bars.<br><em>This is an essential element.</em></td></tr><tr><td>Facets</td><td>Faceting refers to splitting the data into multiple subsets and then displaying plots for the specific subsets in a panel. Such plots are also called small-multiple plots. The&nbsp;<em>facet</em>&nbsp;approach partitions a plot into a matrix of panels. Each panel shows a different subset of the data.<em>This is an optional element.</em></td></tr><tr><td>Statistics</td><td>This refers to representing statistical information about the data, such as mean and variance, to help in understanding the data.<em>This is an optional element.</em></td></tr><tr><td>Coordinates</td><td>This refers to the space on which the data is plotted (E.g., Cartesian coordinates). Most popular graphs such as line and bar charts are drawn using Cartesian coordinates.<br><em>This is an optional element.</em></td></tr><tr><td>Themes</td><td>Themes are used to change the appearance of non-data elements. Themes enable you to design with a particular visual identity using fonts, colours and other design elements.<em>This is an optional element.</em></td></tr></tbody></table><figcaption>7 grammatical elements in ggplot2</figcaption></figure>



<p>It is useful to think of plots made up of layers. A layer can be thought of as consisting of the data, a mapping of data to aesthetics, and a geometry to visually display the data. Sometimes you can also use additional parameters to customize the display.</p>



<p>The grammar of graphics is implemented in R using the ggplot2 package. Essentially we develop plots by layering graphical elements on top of each other and use aesthetic mappings to refine our visualizations.</p>



<p>The following diagram provides an Ariel view of choices we will make while defining each layer in a plot. So, when we are making a plot with&nbsp;<code>ggplot2</code>, we are making choices from these items (Note: These are examples of choices. there are more choices that are not shown.)</p>



<p>In the next few lessons, we will learn about how we can create a variety of meaningful data visualizations using ggplot2 by implementing these 7 layers of grammar of graphics.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot2-grammar-1-1024x819.png" alt="" class="wp-image-27201"/></figure></div>



<p></p>

Grammar of Graphics in ggplot

<p>To learn data visualization with&nbsp;<code>ggplot2</code>&nbsp;in R, we will be making use of various datasets. However, one interesting dataset that we will be using quite a lot in this section is the German Credit dataset.</p>



<p>The German Credit Data contains data on 20 variables and the classification whether an applicant is considered a Good or a Bad credit risk for 1000 loan applicants.</p>



<p>When a bank receives a loan application, based on the applicant’s profile the bank has to make a decision regarding whether to go ahead with the loan approval or not. Two types of risks are associated with the bank’s decision:</p>



<ul><li>If the applicant is a good credit risk, i.e. is likely to repay the loan, then not approving the loan to the person results in a loss of business to the bank</li><li>If the applicant is a bad credit risk, i.e. is not likely to repay the loan, then approving the loan to the person results in a financial loss to the bank</li></ul>



<p>In this section, we will explore the dataset using ggplot2 and create both exploratory as well as explanatory data visualizations. However, later in another course we will also use this dataset to build a predictive credit risk model. To minimize loss from the bank’s perspective, the bank needs a decision rule regarding who to give approval of the loan and who not to. An applicant’s demographic and socio-economic profiles are considered by loan managers before a decision is taken regarding his/her loan application. A predictive model developed on this data is expected to provide a bank manager guidance for making a decision whether to approve a loan to a prospective applicant based on his/her profiles.</p>



<p>You can download the data from the following link:</p>



<div class="wp-block-file"><a href="https://financetrain.com/wp-content/uploads/german-credit.csv">German Credit Dataset</a><a href="https://financetrain.com/wp-content/uploads/german-credit.csv" class="wp-block-file__button" download>Download</a></div>



<h3>Attributes of German Credit Data</h3>



<p>Number of Attributes: 20 (7 numerical, 13 categorical).</p>



<figure class="wp-block-table"><table><tbody><tr><td><strong>Attribute</strong></td><td><strong>Description</strong></td><td><strong>Type</strong></td></tr><tr><td>Status of existing checking account</td><td>A11 : &lt; 0 DM<br>A12 : 0 &lt;= ... &lt; 200 DM<br>A13 : >= 200 DM<br>A14 : No checking account</td><td>Qualitative</td></tr><tr><td>Duration of Credit (in months)</td><td></td><td>Numerical</td></tr><tr><td>Credit history</td><td>A30 : no credits taken/ all credits paid back duly<br>A31 : all credits at this bank paid back duly<br>A32 : existing credits paid back duly till now<br>A33 : delay in paying off in the past<br>A34 : critical account/other credits existing (not at this bank)</td><td>Qualitative</td></tr><tr><td>Purpose of Loan</td><td>A40 : car (new)<br>A41 : car (used)<br>A42 : furniture/equipment<br>A43 : radio/television<br>A44 : domestic appliances<br>A45 : repairs<br>A46 : education<br>A47 : vacation<br>A48 : retraining<br>A49 : business<br>A410 : others</td><td>Qualitative</td></tr><tr><td>Credit amount</td><td></td><td>Numerical</td></tr><tr><td>Savings account/bonds</td><td>A61 : &lt; 100 DM<br>A62 : 100 &lt;= ... &lt; 500 DM<br>A63 : 500 &lt;= ... &lt; 1000 DM<br>A64 : ... >= 1000 DM<br>A65 : unknown/ no savings account</td><td>Qualitative</td></tr><tr><td>Present employment since</td><td>A71 : unemployed<br>A72 : ... &lt; 1 year<br>A73 : 1 &lt;= ... &lt; 4 years<br>A74 : 4 &lt;= ... &lt; 7 years<br>A75 : .. >= 7 years</td><td>Qualitative</td></tr><tr><td>Installment rate (%)</td><td></td><td>Numerical</td></tr><tr><td>Personal status and sex</td><td>A91 : male : divorced/separated <br>A92 : female : divorced/separated/married<br>A93 : male : single<br>A94 : male : married/widowed <br>A95 : female : single</td><td>Qualitative</td></tr><tr><td>Guarantors</td><td>A101 : none<br>A102 : co-applicant<br>A103 : guarantor</td><td>Qualitative</td></tr><tr><td>Present residence since</td><td></td><td>Numerical</td></tr><tr><td>Most valuable available asset</td><td>A121 : real estate<br>A122 : if not <br>A121 : building society savings agreement/ life insurance<br>A123 : if not A121/A122 : car or other, not in attribute 6<br>A124 : unknown / no property</td><td>Qualitative</td></tr><tr><td>Age in years</td><td></td><td>Numerical</td></tr><tr><td>Concurrent Credits</td><td>A141 : bank<br>A142 : stores<br>A143 : none</td><td>Qualitative</td></tr><tr><td>Type of housing</td><td>A151 : rent<br>A152 : own<br>A153 : for free</td><td>Qualitative</td></tr><tr><td>Number of existing credits at this bank</td><td></td><td>Numerical</td></tr><tr><td>Job</td><td>A171 : unemployed/ unskilled - non-resident<br>A172 : unskilled - resident<br>A173 : skilled employee / official<br>A174 : management/ self-employed/ highly qualified employee/ officer</td><td>Qualitative</td></tr><tr><td>No of dependents</td><td></td><td>Numerical</td></tr><tr><td>Telephone</td><td>A191 : none<br>A192 : yes, registered under the customers name</td><td>Qualitative</td></tr><tr><td>Foreign Worker</td><td>A201 : yes<br>A202 : no</td><td>Qualitative</td></tr><tr><td>Loan Quality</td><td>1 : Bad loan<br>2 : Good loan</td><td>Qualitative</td></tr></tbody></table><figcaption>Attributes of German Credit Dataset</figcaption></figure>



<p>While we use this data to learn the techniques of data visualizations, we will also be learning other important principles of data science, specially the process of data cleaning. We will work on this data to make it suitable for our analysis and to make visualizations meaningful.</p>



<h3>Data Import and Basic Manipulation</h3>



<p>Now that we have the data in CSV format, we will first import it into R as a data frame. I first placed the CSV file in a folder of my choice, then updated my working directory to the folder where the data file is stored. Then I used the <code>read.csv()</code> command to import the CSV data into a data frame called <code>df</code>.</p>



<pre class="wp-block-code"><code>> setwd("C:/Users/FT/Dropbox/FinanceTrain/Courses/Data")
> getwd()
&#91;1] "C:/Users/FT/Dropbox/FinanceTrain/Courses/Data"
> df &lt;- read.csv("german-credit.csv")</code></pre>



<pre class="wp-block-preformatted">The data is now loaded into our R session in <code>df</code> dataframe. We can inspect the structure of the dataframe using the <code>str()</code> function.</pre>



<pre class="wp-block-code"><code>> str(df)
'data.frame':    1000 obs. of  21 variables:
 $ Status.of.existing.checking.account    : Factor w/ 4 levels "A11","A12","A13",..: 1 2 4 1 1 4 4 2 4 2 ...
 $ Duration.of.Credit..in.months.         : int  6 48 12 42 24 36 24 36 12 30 ...
 $ Credit.history                         : Factor w/ 5 levels "A30","A31","A32",..: 5 3 5 3 4 3 3 3 3 5 ...
 $ Purpose.of.Loan                        : Factor w/ 10 levels "A40","A41","A410",..: 5 5 8 4 1 8 4 2 5 1 ...
 $ Credit.amount                          : int  1169 5951 2096 7882 4870 9055 2835 6948 3059 5234 ...
 $ Savings.account.bonds                  : Factor w/ 5 levels "A61","A62","A63",..: 5 1 1 1 1 5 3 1 4 1 ...
 $ Present.employment.since               : Factor w/ 5 levels "A71","A72","A73",..: 5 3 4 4 3 3 5 3 4 1 ...
 $ Installment.rate....                   : int  4 2 2 2 3 2 3 2 2 4 ...
 $ Personal.status.and.sex                : Factor w/ 4 levels "A91","A92","A93",..: 3 2 3 3 3 3 3 3 1 4 ...
 $ Guarantors                             : Factor w/ 3 levels "A101","A102",..: 1 1 1 3 1 1 1 1 1 1 ...
 $ Present.residence.since                : int  4 2 3 4 4 4 4 2 4 2 ...
 $ Most.valuable.available.asset          : Factor w/ 4 levels "A121","A122",..: 1 1 1 2 4 4 2 3 1 3 ...
 $ Age.in.years                           : int  67 22 49 45 53 35 53 35 61 28 ...
 $ Concurrent.Credits                     : Factor w/ 3 levels "A141","A142",..: 3 3 3 3 3 3 3 3 3 3 ...
 $ Type.of.housing                        : Factor w/ 3 levels "A151","A152",..: 2 2 2 3 3 3 2 1 2 2 ...
 $ Number.of.existing.credits.at.this.bank: int  2 1 1 1 2 1 1 1 1 2 ...
 $ Job                                    : Factor w/ 4 levels "A171","A172",..: 3 3 2 3 3 2 3 4 2 4 ...
 $ No.of.dependents                       : int  1 1 2 2 2 2 1 1 1 1 ...
 $ Telephone                              : Factor w/ 2 levels "A191","A192": 2 1 1 1 1 2 1 2 1 1 ...
 $ Foreign.Worker                         : Factor w/ 2 levels "A201","A202": 1 1 1 1 1 1 1 1 1 1 ...
 $ Loan.Quality                           : int  1 2 1 1 2 1 1 1 1 2 ...</code></pre>



<p>A few observations about the data:</p>



<ul><li>There are 1000 observations of 21 variables.</li><li>R keeps the numeric data as it is (int)</li><li>R converts the qualitative data in factors (categories). These are also called levels. Internally, R stores the integer values 1, 2, and 3, and maps the character strings (in alphabetical order, unless I reorder) to these values. For example, the variable&nbsp;<code>Foreign.Worker</code>&nbsp;has two levels, namely, A201 and A202 which correspond to 'Yes' and No' in our data. Internally data is stored as 1 and 2.</li></ul>



<h3>Relabeling the Factor Levels</h3>



<p>Our factors are labeled using some internal codes which are a bit difficult to remember. They are also not very useful data visualizations. For example, in case of Foreign.Worker, the levels A201 and A202 not very intuitive even though we know that they mean 'Yes and 'No' to signify whether it is a foreign worker or not.</p>



<p>In such situation we can easily rename the levels by supplying a new vector of labels. The following example shows changing the level names from A201 and A202 to Yes and No.</p>



<pre class="wp-block-code"><code>> levels(df$Foreign.Worker) &lt;- c('Yes','No')
> str(df)
'data.frame':    1000 obs. of  21 variables:
 $ Status.of.existing.checking.account    : Factor w/ 4 levels "A11","A12","A13",..: 1 2 4 1 1 4 4 2 4 2 ...
 $ Duration.of.Credit..in.months.         : int  6 48 12 42 24 36 24 36 12 30 ...
 $ Credit.history                         : Factor w/ 5 levels "A30","A31","A32",..: 5 3 5 3 4 3 3 3 3 5 ...
 $ Purpose.of.Loan                        : Factor w/ 10 levels "A40","A41","A410",..: 5 5 8 4 1 8 4 2 5 1 ...
 $ Credit.amount                          : int  1169 5951 2096 7882 4870 9055 2835 6948 3059 5234 ...
 $ Savings.account.bonds                  : Factor w/ 5 levels "A61","A62","A63",..: 5 1 1 1 1 5 3 1 4 1 ...
 $ Present.employment.since               : Factor w/ 5 levels "A71","A72","A73",..: 5 3 4 4 3 3 5 3 4 1 ...
 $ Installment.rate....                   : int  4 2 2 2 3 2 3 2 2 4 ...
 $ Personal.status.and.sex                : Factor w/ 4 levels "A91","A92","A93",..: 3 2 3 3 3 3 3 3 1 4 ...
 $ Guarantors                             : Factor w/ 3 levels "A101","A102",..: 1 1 1 3 1 1 1 1 1 1 ...
 $ Present.residence.since                : int  4 2 3 4 4 4 4 2 4 2 ...
 $ Most.valuable.available.asset          : Factor w/ 4 levels "A121","A122",..: 1 1 1 2 4 4 2 3 1 3 ...
 $ Age.in.years                           : int  67 22 49 45 53 35 53 35 61 28 ...
 $ Concurrent.Credits                     : Factor w/ 3 levels "A141","A142",..: 3 3 3 3 3 3 3 3 3 3 ...
 $ Type.of.housing                        : Factor w/ 3 levels "A151","A152",..: 2 2 2 3 3 3 2 1 2 2 ...
 $ Number.of.existing.credits.at.this.bank: int  2 1 1 1 2 1 1 1 1 2 ...
 $ Job                                    : Factor w/ 4 levels "A171","A172",..: 3 3 2 3 3 2 3 4 2 4 ...
 $ No.of.dependents                       : int  1 1 2 2 2 2 1 1 1 1 ...
 $ Telephone                              : Factor w/ 2 levels "A191","A192": 2 1 1 1 1 2 1 2 1 1 ...
 $ Foreign.Worker                         : Factor w/ 2 levels "Yes","No": 1 1 1 1 1 1 1 1 1 1 ...
 $ Loan.Quality                           : int  1 2 1 1 2 1 1 1 1 2 ...</code></pre>



<p>We can also check the levels of any variable using the <code>levels()</code> function as shown below:</p>



<pre class="wp-block-code"><code>> levels(df$Foreign.Worker)
&#91;1] "Yes" "No"</code></pre>



<p>Similarly, we will update the level names for all the qualitative variables in the dataframe to suite our requirements. The following script does that.</p>



<pre class="wp-block-code"><code>levels(df$Status.of.existing.checking.account) &lt;- c('&lt; 0 DM','0 - 200 DM', '>= 200 DM', 'No checking account')
levels(df$Credit.history) &lt;- c('No Credits Taken','All Credit Paid', 'Existing Credit Paid','Delay in Payment','Critical Account')
levels(df$Purpose.of.Loan) &lt;- c('car (new)', 'car (used)', 'furniture/equipment', 'radio/television', 'domestic appliances', 'repairs', 'education', 'vacation', 'retraining', 'business', 'others')
levels(df$Savings.account.bonds) &lt;- c('&lt;  100 DM', '100 - 500 DM', '500 - 1000 DM', '>= 1000 DM', 'No Savings Account')
levels(df$Present.employment.since) &lt;- c('unemployed', '&lt; 1 year', '1 - 4 years', '4 - 7 years', '>= 7 years')
levels(df$Personal.status.and.sex) &lt;- c('male : divorced/separated', 'female : divorced/separated/married', 'male : single', 'male : married/widowed', 'female : single')
levels(df$Guarantors) &lt;- c('none', 'co-applicant', 'guarantor')
levels(df$Most.valuable.available.asset) &lt;- c('real estate', 'savings agreement/life insurance', 'car or other', 'unknown / no property')
levels(df$Concurrent.Credits) &lt;- c('bank', 'stores', 'none')
levels(df$Type.of.housing) &lt;- c('rent', 'own', 'for free')
levels(df$Job) &lt;- c('unemployed/ unskilled - non-resident', 'unskilled - resident', 'skilled employee / official', 'management/ self-employed')
levels(df$Telephone) &lt;- c('No','Yes')
levels(df$Foreign.Worker) &lt;- c('Yes','No')</code></pre>



<p>The above script will update the level names for all the variables as per the labels provided for us. Let's inspect the structure again.</p>



<pre class="wp-block-code"><code>> str(df)
'data.frame':    1000 obs. of  21 variables:
 $ Status.of.existing.checking.account    : Factor w/ 4 levels "&lt; 0 DM","0 - 200 DM",..: 1 2 4 1 1 4 4 2 4 2 ...
 $ Duration.of.Credit..in.months.         : int  6 48 12 42 24 36 24 36 12 30 ...
 $ Credit.history                         : Factor w/ 5 levels "No Credits Taken",..: 5 3 5 3 4 3 3 3 3 5 ...
 $ Purpose.of.Loan                        : Factor w/ 11 levels "car (new)","car (used)",..: 5 5 8 4 1 8 4 2 5 1 ...
 $ Credit.amount                          : int  1169 5951 2096 7882 4870 9055 2835 6948 3059 5234 ...
 $ Savings.account.bonds                  : Factor w/ 5 levels "&lt;  100 DM","100 - 500 DM",..: 5 1 1 1 1 5 3 1 4 1 ...
 $ Present.employment.since               : Factor w/ 5 levels "unemployed","&lt; 1 year",..: 5 3 4 4 3 3 5 3 4 1 ...
 $ Installment.rate....                   : int  4 2 2 2 3 2 3 2 2 4 ...
 $ Personal.status.and.sex                : Factor w/ 5 levels "male : divorced/separated",..: 3 2 3 3 3 3 3 3 1 4 ...
 $ Guarantors                             : Factor w/ 3 levels "none","co-applicant",..: 1 1 1 3 1 1 1 1 1 1 ...
 $ Present.residence.since                : int  4 2 3 4 4 4 4 2 4 2 ...
 $ Most.valuable.available.asset          : Factor w/ 4 levels "real estate",..: 1 1 1 2 4 4 2 3 1 3 ...
 $ Age.in.years                           : int  67 22 49 45 53 35 53 35 61 28 ...
 $ Concurrent.Credits                     : Factor w/ 3 levels "bank","stores",..: 3 3 3 3 3 3 3 3 3 3 ...
 $ Type.of.housing                        : Factor w/ 3 levels "rent","own","for free": 2 2 2 3 3 3 2 1 2 2 ...
 $ Number.of.existing.credits.at.this.bank: int  2 1 1 1 2 1 1 1 1 2 ...
 $ Job                                    : Factor w/ 4 levels "unemployed/ unskilled - non-resident",..: 3 3 2 3 3 2 3 4 2 4 ...
 $ No.of.dependents                       : int  1 1 2 2 2 2 1 1 1 1 ...
 $ Telephone                              : Factor w/ 2 levels "No","Yes": 2 1 1 1 1 2 1 2 1 1 ...
 $ Foreign.Worker                         : Factor w/ 2 levels "Yes","No": 1 1 1 1 1 1 1 1 1 1 ...
 $ Loan.Quality                           : int  1 2 1 1 2 1 1 1 1 2 ...
></code></pre>



<p>Our data is now ready for data visualization.</p>

Data Import and Basic Manipulation in R - German Credit Dataset

<p>In ggplot2 package, we use the&nbsp;<code>ggplot()</code>&nbsp;function to create a fully customized data visualization. We still have the German credit data loaded in the dataframe&nbsp;<code>df</code>.</p>



<p>We will start by plotting a simple scatter graph that plots the duration of credit on x-axis and the amount of credit on y-axis. We can do so using the following command.</p>



<pre class="wp-block-code"><code>> ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount))

or 

g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount))</code></pre>



<p>The process of creating a graph starts with the&nbsp;<code>ggplot()</code>&nbsp;function. Note that we can either directly issue the command which will print the graph or we can create an object by assigning the function to a variable. In the 2nd example above, we have created an R object called g that stores the graph object.</p>



<p>The above command uses only the first two levels of the grammar of graphics, i.e., data and aesthetics. If we print this, it will only print an empty graph, as shown below:</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot1.png" alt="" class="wp-image-27215"/></figure></div>



<p>We use the third level, that is, geom (geometric object) to print the actual data on the graph. Geometric objects are the actual marks we put on a plot.</p>



<ul><li>points (geom_point, for scatter plots, dot plots, etc)</li><li>lines (geom_line, for time series, trend lines, etc)</li><li>boxplot (geom_boxplot, for, well, boxplots!)</li></ul>



<p>We use&nbsp;<code>geom_points</code>&nbsp;to create scatter plots,&nbsp;<code>geom_bar</code>&nbsp;for bar chart and so on. A plot must have at least one geom; there is no upper limit. You can add a geom to a plot using the&nbsp;<code>+</code>&nbsp;operator.</p>



<p>You can get a list of available geometric objects using the code below:</p>



<pre class="wp-block-code"><code>help.search("geom_", package = "ggplot2")</code></pre>



<h3>Points (Scatterplot)</h3>



<p>We now know the data that we want to plot, the aesthetics, and the geometric object that we want to create. We can use this information to complete our scatter plot.</p>



<p>We will add the points geom to our graph object <code>g</code>, as shown below:</p>



<pre class="wp-block-code"><code>> g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount))
> g + geom_point()</code></pre>



<p>The scatter plot will be created as shown below:</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot2.png" alt="" class="wp-image-27216"/></figure></div>



<p>The above scatter plot shows the relationship between the duration of credit in months and the amount loan. Each point represents a loan. In our dataset, we have 1000 loans and we also know that each loan has either defaulted or not. This is represented in the data point&nbsp;<code>Loan.Quality</code>. We can use this variable to improve this graph, for example, by coloring the points based on the loan quality.</p>



<p>However, there is one small problem here. The variable <code>Loan.Quality</code> is of type integer (1 for Bad Loan, and 2 for Good Loan). So, to use to to categorize data, we can convert it to Factor and then add levels.</p>



<pre class="wp-block-code"><code>#Convert Loan.Quality to Factor
df$Loan.Quality &lt;- as.factor(df$Loan.Quality)
#Add levels for data points
levels(df$Loan.Quality) &lt;- c("Bad Loan", "Good Loan")</code></pre>



<p>We can now use this variable to color the points. For this we will add the parameter color to aesthetics.</p>



<pre class="wp-block-code"><code>> g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount))
> g+geom_point(aes(color=Loan.Quality))</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot3.png" alt="" class="wp-image-27217"/></figure></div>



<p>As you can see, the points in the scatter plot are of two colors, red for bad loans, and blue for good loans. ggplot2 has automatically selected the colors and also added the legend for convenience.</p>

Create ggplot Graph with German Credit Data in R

<p>Faceting refers to splitting the data into multiple subsets and then displaying plots for the specific subsets in a panel. Such plots are also called small-multiple plots. The facet approach partitions a plot into a matrix of panels. Each panel shows a different subset of the data.</p>



<p>In our scatter plot we grouped the data into bad and good loans using colors. We can use the faceting concept to split the entire graph into two separate panels one for each, good and bad loans. To do so we will use the&nbsp;<code>facet_grid(. ~ ___)</code>&nbsp;layer.&nbsp;<code>facet_grid</code>&nbsp;forms a matrix of panels defined by row and column faceting variables.&nbsp;<code>facet_grid(x ~ y)</code>&nbsp;will display x*y plots even if some plots are empty.</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount))
g+geom_point()+geom_smooth(se=FALSE)+
  facet_grid(. ~ Loan.Quality)</code></pre>



<p>We split our data only by columns. That's why we specify the rows by a simple. The columns are split by Loan.Quality. Since this variable has only two values, the graph will be split into two columns.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplotf1-1.png" alt="" class="wp-image-27233"/></figure></div>



<p>Faceting can be specially useful when we want to bring another variable into our plot. For example in this case, we can now add the Job variable to see how the type of job of the borrower impacts the credit.</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount, color=Job))
g+geom_point()+geom_smooth(se=FALSE,method='lm')+
  facet_grid(. ~ Loan.Quality)</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplotf2.png" alt="" class="wp-image-27234"/></figure></div>



<p><code>ggplot2</code> is beautiful and to slice the data in a different way is extremely easy. In the following graph, we exchange the two parameters <code>Loan.Quality</code> and <code>Job</code>.</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount, color=Loan.Quality))
g+geom_point()+geom_smooth(se=FALSE,method='lm')+
  facet_grid(. ~ Job)</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplotf3.png" alt="" class="wp-image-27235"/></figure></div>



<p></p>

Splitting Plots with Facets in ggplots

<h3>ggplot2 - Aestheitics</h3>



<p>Generally when we talk about aesthetics, we talk about the attributes of a chart such as the color, size and shape. However, in ggplot2, it is not just about how something looks but also about how a variable is mapped it to. In our earlier examples, we mapped the Loan.Quality to Color. As a part of the Aesthetics layer, we map variables to aesthetics. This includes a variety of things such as the x-position, y-position, color, fill and so on. So, when we want to plot Credit Amount on y-axis, we are essentially mapping the variable Credit.Amount to y axis.</p>



<p>The following is a list of various aesthetics that we can specify.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-aes-1-978x1024.png" alt="" class="wp-image-27246"/></figure></div>



<p>Let's take an example and try out these aesthetics. We will use the German Credit Data and create a scatter plot with custom aesthetics.</p>



<ul><li>Map Credit.amount on x-axis</li><li>Map Duration.of.Credit..in.months. on y-axis</li><li>Change shape to 21 (filled circle with outline). Use ?shape to learn about different shapes available.</li><li>Map the Loan.Quality to fill</li><li>Change the shape size to 3</li><li>Reduce alpha to 0.5</li><li>Add a plot title - "Loan Amount Vs Duration"</li><li>Change x-axis label to "Loan Amount"</li><li>Change y-axis label to "Duration in Months"</li></ul>



<p>I suggest that you try doing this yourself in your R environment and if things don't work out, use the code provided below:</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Credit.amount,y=Duration.of.Credit..in.months.,fill=Loan.Quality))
g+geom_point(shape=21,size=3,alpha=0.5)</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-aes2.png" alt="" class="wp-image-27247"/></figure></div>



<p>Let's plot one more graph and we will also make use of facets to split the data into multiple plots.</p>



<ul><li>Map Job to x-axis</li><li>Map Credit.amount to y-axis</li><li>May Job to color so that each job category has a different color.</li><li>Add jitter to the data. Jittering refers to purposely adding noise to your data. This will be helpful here because on x-axis we have a category. If we simply plot the points they will all be in one line</li><li>Change alpha to 0.5</li><li>Change size to 3</li></ul>



<p>The following code achieves this:</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Job,y=Credit.amount,color=Job))
g+geom_jitter(size=2,alpha=0.5)+
labs(title ="Loan Amount by Job", x = "Job", y = "Loan Amount")</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-aes3.png" alt="" class="wp-image-27248"/></figure></div>



<p>The way ggplot2 is designed is that you can customize almost anything in the chart and achieve exactly what you want.&nbsp;</p>



<h3>ggplot2 - Position Adjustments</h3>



<p><code>ggplot2</code>&nbsp;allows us to adjust the position of each geom. To do so, we simply have to specify the desired position to the position argument of the geom function. In the previous lesson, we saw jittering, which is an example of position adjustment of continuous data.</p>



<p>We have the following position adjustments available:</p>



<ul><li><code>position_identity</code>&nbsp;- This is the default for most geoms. This just means don't adjust position. So we are telling ggplot2 to plot the data as it is.</li><li><code>position_jitter</code>&nbsp;- This allows us to add noise to the plot which may be hard to read because of multiple overlapping points. We can specify width and height as arguments -&nbsp;<code>position_jitter(width = NULL, height = NULL)</code></li><li><code>position_dodge</code>&nbsp;- Dodging preserves the vertical position of a geom while adjusting the horizontal position. Format:&nbsp;<code>position_dodge(width = NULL, preserve = c("total", "single"))</code>.</li><li><code>position_stack</code>&nbsp;- Stacks bars on to of each other. This is the default of&nbsp;<code>geom_bar</code>&nbsp;and&nbsp;<code>geom_area</code></li><li><code>position_fill</code>&nbsp;- stacks bars and standardizes each stack to have constant height</li></ul>



<h4>geom_bar in ggplot2</h4>



<p>Let's learn about position adjustments using&nbsp;<code>geom_bar</code>&nbsp;in ggplot2. We will use our German Credit dataset.</p>



<h4>Simple Bar Chart</h4>



<p>We will start by plotting a simple bar chart with the borrower's Credit History on x-axis and the amount of loan taken on y-axis.</p>



<pre class="wp-block-code"><code>p &lt;- ggplot(data=df, aes(x=Credit.history, y=Credit.amount)) +
  geom_bar(stat="identity")
p</code></pre>



<p>This will produce the following chart:</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-pos1.png" alt="" class="wp-image-27250"/></figure></div>



<h3>Add Colors to Bars</h3>



<p>To add colors to the bars, we can supply the x dimension to the fill argument (<code>fill = Credit.history</code>). ggplot2 will assign a color to each value in Credit.history and fill the bars with that color.</p>



<pre class="wp-block-code"><code>p &lt;- ggplot(data=df, aes(x=Credit.history, y=Credit.amount, fill=Credit.history )) +
  geom_bar(stat="identity")
p</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-pos2.png" alt="" class="wp-image-27251"/></figure></div>



<h3>Stacked Bars</h3>



<p>We can also group the data by a third variable such as <code>Loan.Quality</code>. By default it will split the data by that variable and plot a stacked bar chart. This is done by mapping the variable <code>Loan.Quality</code> to the fill scale. <code>position_stack</code> is the default argument for geom_bar</p>



<pre class="wp-block-code"><code>p &lt;- ggplot(data=df, aes(x=Credit.history, y=Credit.amount, fill=Loan.Quality )) +
  geom_bar(stat="identity")
p</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-pos3.png" alt="" class="wp-image-27252"/></figure></div>



<h3>Dodging</h3>



<p>If we did not want the bars to be stacked we can use <code>position_dodge</code> which will preserve the vertical position of the geom while adjusting the horizontal position.</p>



<pre class="wp-block-code"><code>p &lt;- ggplot(data=df, aes(x=Credit.history, y=Credit.amount, fill=Loan.Quality )) +
  geom_bar(stat="identity",position=position_dodge())
p</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-pos4.png" alt="" class="wp-image-27253"/></figure></div>



<h3>100% Stacked Bar</h3>



<p>We can use the argument position_fill to change it into a 100% stacked bar chart which is useful for relative comparison.</p>



<pre class="wp-block-code"><code>p &lt;- ggplot(data=df, aes(x=Credit.history, y=Credit.amount, fill=Loan.Quality )) +
  geom_bar(stat="identity",position=position_fill())
p</code></pre>



<figure class="wp-block-image size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-pos5.png" alt="" class="wp-image-27254"/></figure>



<h3>Modifying Scales</h3>



<p>In <code>ggplot2</code>, each aesthetic is a scale which we map our data on to. So, color is just a scale just like x and y axis. We can access and modify each scale using the <code>"scale_"</code> functions. For example, we can use <code>scale_fill_manual</code> to manually specify the colors to be used in the plot as shown in the plot below:</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-pos6.png" alt="" class="wp-image-27255"/></figure></div>



<p></p>

ggplot2 -  Chart Aesthetics and Position Adjustments in R

<p>Apart from scatter and bar charts, another popular type of chart that is frequently used in financial analysis is the line chart. In this lesson we will learn about how to create a line chart using&nbsp;<code>ggplot2</code>.</p>



<p>Line charts are best suited for time-series data with time/date on the x-axis and some metrics or data on the y-axis. For our example, we will use stock prices data.</p>



<h3>Import Data</h3>



<p>We have created a CSV file that contains the daily stock prices of the top 5 tech stocks in the US for the past 1 year. </p>



<div class="wp-block-file"><a href="https://financetrain.com/wp-content/uploads/stock_data-1.csv">stock_data</a><a href="https://financetrain.com/wp-content/uploads/stock_data-1.csv" class="wp-block-file__button" download>Download</a></div>



<p>Let's first import the data into R as a data frame (called <code>stock_prices</code>). We can inspect the data frame using the <code>str()</code> function.</p>



<pre class="wp-block-code"><code>> stock_prices &lt;-read.csv("stock_data.csv")
> str(stock_prices)
'data.frame':    251 obs. of  6 variables:
 $ Date : Factor w/ 251 levels "1/10/2017","1/11/2017",..: 179 180 181 182 188 204 205 206 207 189 ...
 $ AAPL : num  92 93.6 94.4 95.6 95.9 ...
 $ MSFT : num  48.4 49.4 50.5 51.2 51.2 ...
 $ GOOGL: num  681 691 695 704 710 ...
 $ IBM  : num  144 146 148 152 152 ...
 $ INTC : num  30.7 31.2 31.9 32.8 32.8 ...
></code></pre>



<p>One problem with the above data is that the data in date field is taken as Factors instead of Dates. Before proceeding, we will convert this field into a date field.</p>



<pre class="wp-block-code"><code>> stock_prices$Date&lt;-as.Date(as.character(stock_prices$Date), format = '%m/%d/%Y')
> str(stock_prices)
'data.frame':    251 obs. of  6 variables:
 $ Date : Date, format: "2016-06-27" "2016-06-28" ...
 $ AAPL : num  92 93.6 94.4 95.6 95.9 ...
 $ MSFT : num  48.4 49.4 50.5 51.2 51.2 ...
 $ GOOGL: num  681 691 695 704 710 ...
 $ IBM  : num  144 146 148 152 152 ...
 $ INTC : num  30.7 31.2 31.9 32.8 32.8 ...
></code></pre>



<p>Now you can see that the format for the <code>Date</code> variable is Date.</p>



<h3>Simple Line Chart</h3>



<p>Let's plot our first line chart with date on x-axis and Apple stock prices on y-axis.</p>



<pre class="wp-block-code"><code>ggplot(stock_prices, aes(x = Date, y = AAPL)) +
  geom_line()</code></pre>



<p><code>ggplot</code> will automatically adjust both x and y-axis scales according to the data.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-line1.png" alt="" class="wp-image-27309"/><figcaption>Line Chart in ggplot2</figcaption></figure></div>



<h3>Customizing Line Aesthetics</h3>



<p>You can customize various aspects of a line in the plot such as the line width, line color, and line type.</p>



<ul><li><code>linetype</code>: This parameter specifies the line type. The common options are "solid", "dashed", "dotted", "dotdash", "longdash", "twodash".</li><li><code>size</code>: This parameter specifies the thinkness of the line specified as a number. Default is 0.</li><li><code>colour</code>: This parameter is used to specify the color of the line.</li></ul>



<p>In the following code, we change the <code>linetype</code> to <code>dashed</code>, increase the size to 1, and the colour of the line to blue.</p>



<pre class="wp-block-code"><code>ggplot(stock_prices, aes(x = Date, y = AAPL)) +
  geom_line(colour="blue", linetype="dashed", size=1)</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-line2.png" alt="" class="wp-image-27310"/></figure></div>



<h3>Plotting Multiple Lines (Cleaning Data)</h3>



<p>Let's say we want to plot the prices of all the stocks in a single plot with each line representing one stock. One way to do this is to use <code>geom_line()</code> multiple times to add each additional line.</p>



<pre class="wp-block-code"><code>ggplot(stock_prices, aes(x = Date, y = AAPL)) +
  geom_line()+
  geom_line(aes(x = Date, y = GOOGL))</code></pre>



<p>However, this is not the correct method.</p>



<p>The correct way to handle this scenario is to convert the data into long form. This is also called tidy data where each column is a variable and each row is an observation. In our example data, we would want to have a separate column called 'Symbol' which will hold the stock symbol, and a new column called Price which will hold the price for that symbol.</p>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-line3.png" alt="" class="wp-image-27311"/></figure></div>



<p>Once we have data in this format, we can plot the line chart and group the data by Symbol to split it into multiple lines.</p>



<p>To convert the data into long-form, we can use the <code>tidyr</code> package. (We will have an elaborate course on cleaning data in R).</p>



<pre class="wp-block-code"><code>#install and load tidyr package
install.packages("tidyverse")
library(tidyr)</code></pre>



<p><code>tidyverse</code>&nbsp;installs many packages related to cleaning data.</p>



<p>We use the&nbsp;<code>gather()</code>&nbsp;function of the&nbsp;<code>tidyr</code>&nbsp;package to move data from&nbsp;<code>stock_prices</code>&nbsp;dataframe to a new dataframe called&nbsp;<code>stock_prices.tidy</code>. This data frame should have three columns:&nbsp;<code>Date</code>,&nbsp;<code>Symbol</code>&nbsp;and&nbsp;<code>Prices</code>.</p>



<p><code>gather()</code> takes four arguments: the original data frame (stock_prices), the key column (Symbol), the value column (Prices) and the name of the grouping variable with a minus in front (-Date).</p>



<pre class="wp-block-code"><code>> stock_prices.tidy &lt;- gather(stock_prices,Symbol, Prices, -Date)</code></pre>



<p>You can now inspect the dataframe.</p>



<pre class="wp-block-code"><code>> str(stock_prices.tidy)
'data.frame':    1255 obs. of  3 variables:
 $ Date  : Date, format: "2016-06-27" ...
 $ Symbol: chr  "AAPL" "AAPL" "AAPL" "AAPL" ...
 $ Prices: num  92 93.6 94.4 95.6 95.9 ...</code></pre>



<h3>Plotting Multiple Lines</h3>



<p>We can now easily plot multiple lines with this new dataframe.</p>



<pre class="wp-block-code"><code>ggplot(stock_prices.tidy, aes(x = Date, y = Prices,color = Symbol)) +
  geom_line()</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggolot-line4.png" alt="" class="wp-image-27312"/></figure></div>



<h3>Subsetting Data</h3>



<p>Let's say we don't want to exclude <code>GOOGL</code> data from the plot. To do so, we can subset the data using the <code>subset</code> function and store the resulting dataset in a new data frame. Then we can use the new data frame to create the line chart.</p>



<pre class="wp-block-code"><code>newdata &lt;- subset(stock_prices.tidy,Symbol!='GOOGL')
ggplot(newdata, aes(x = Date, y = Prices,color = Symbol)) +
  geom_line(size=1)</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-line5.png" alt="" class="wp-image-27313"/></figure></div>



<p><strong>Note about Dates:</strong> In the <code>as.Date()</code> function, the format specifies the format of the input. In this example, the date in the CSV file is structured as "1/10/2017","1/11/2017", i.e., Month, Day and the Year. That's why the format is specified as %m/%d/%y. The output of the <code>as.Date()</code> function is the date, which is printed as 2016/10/12 (YYYY/MM/DD)</p>

Creating a Line Chart in ggplot 2 in R

<p>R has become very popular not just because of its data visualization capabilities but also its ability to combine data visualizations with statistical information. The statistics layer helps us represent statistical information about the data, such as mean and variance, to help in understanding the data.</p>



<p>We earlier learned about geoms which stand for "geometric objects." These are the core elements that you see on the plot, and includes objects such as points, lines, areas, and curves. Stats stand for "statistical transformations." These help us summarize the data in different ways such as counting observations, creating a linear regression line that best fits the data, or adding a confidence interval to the regression line.</p>



<p>All geoms have a default stat. For example, when we create a bar chart using <code>geom_bars</code>, the default stat is <code>stat_count</code>, which counts the number of rows to create the bar chart. You can add statistics layers to the plot using the <code>stat_</code> functions.</p>



<figure class="wp-block-table"><table><tbody><tr><th>Stat</th><th>Description</th></tr><tr><td>bin</td><td>Divide continuous range into bins, and count number of points in each</td></tr><tr><td>boxplot</td><td>Compute statistics necessary for boxplot</td></tr><tr><td>contour</td><td>Calculate contour lines</td></tr><tr><td>density</td><td>Compute 1d density estimate</td></tr><tr><td>identity</td><td>Identity transformation,f(x)=x</td></tr><tr><td>jitter</td><td>Jitter values by adding small random value</td></tr><tr><td>qq</td><td>Calculate values for quantile-quantile plot</td></tr><tr><td>quantile</td><td>Quantile regression</td></tr><tr><td>smooth</td><td>Smoothed conditional mean of y given x</td></tr><tr><td>summary</td><td>Aggregate values of y for given x</td></tr><tr><td>unique</td><td>Remove duplicated observations</td></tr></tbody></table></figure>



<p>Let's look at some of these statistics.</p>



<h3>Smoothing</h3>



<p>One of the most common statistical transformations we add to our plots is a smoothing line. When we plot a scatter chart, we can add a smoothing line using the&nbsp;<code>geom_smooth()</code>, which in turn uses stat_smooth to plot the smooth curve.</p>



<p>Let's use our stock_prices dataset and create a scatter plot with AAPL prices on x-axis and GOOGL prices on y-axis. We will add a smooth line to the scatter plot (LOESS is the default) with <code>geom_smooth()</code>.</p>



<pre class="wp-block-code"><code>ggplot(stock_prices,aes(x=AAPL,y=GOOGL))+
  geom_point()+
  geom_smooth()</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-stat1.png" alt="" class="wp-image-27370"/></figure></div>



<p>We will now change the smoothing to linear model. The default formula is <code>y ~ x</code>. We change the model by specifying <code>method="lm"</code>.</p>



<pre class="wp-block-code"><code>ggplot(stock_prices,aes(x=AAPL,y=GOOGL))+
  geom_point()+
  geom_smooth(method="lm")</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-stat2.png" alt="" class="wp-image-27371"/></figure></div>



<p>The smoothing line shows 95% confidence interval bands by default. We can hide the bands by adding the argument <code>se=FALSE</code>.</p>



<pre class="wp-block-code"><code>ggplot(stock_prices,aes(x=AAPL,y=GOOGL))+
  geom_point()+
  geom_smooth(method="lm",se=FALSE)</code></pre>



<figure class="wp-block-image size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-stat3.png" alt="" class="wp-image-27372"/></figure>



<p>Suppose we wanted to plot just the model and not the actual points. we can do so using <code>geom_smooth()</code> or <code>stat_smooth()</code></p>



<pre class="wp-block-code"><code>ggplot(stock_prices,aes(x=AAPL,y=GOOGL))+
  stat_smooth(method="lm",se=FALSE)</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-stat4.png" alt="" class="wp-image-27373"/></figure></div>



<h3>Modifying stat_smooth</h3>



<p>We can modify various aspects of&nbsp;<code>stat_smooth</code>. For example,&nbsp;<code>span</code>&nbsp;can be used to control the amount of smoothing for the default loess smoother. Smaller numbers produce wigglier lines, larger numbers produce smoother lines.</p>



<p>In the following example, we set the span to 0.3 and also change the line color to red.</p>



<pre class="wp-block-code"><code>ggplot(stock_prices,aes(x=AAPL,y=GOOGL))+
  geom_point()+
  stat_smooth(se=FALSE,col="red",span=0.3)</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-stat5.png" alt="" class="wp-image-27374"/></figure></div>



<p></p>

Add a Statistical Layer on Line Chart in ggplot2

<p><code>stat_summary</code>&nbsp;is a unique statistical function and allows a lot of flexibility in terms of specifying the summary. Using this, you can add a variety of summary on your plots. For example, in a bar chart, you can plot the bars based on a summary statistic such as mean or median. Similarly,&nbsp;<code>stat_summary()</code>&nbsp;can be used to add mean/median points to a dot plot.</p>



<p><code>stat_summary()</code>&nbsp;takes a few different arguments.</p>



<ul><li><code>fun.y</code>: A function to produce y aesthetics</li><li><code>fun.ymax</code>: A function to produce ymax aesthetics</li><li><code>fun.ymin</code>: A function to produce ymin aesthetics</li><li><code>fun.data</code>: A function to produce a named vector of aesthetics</li></ul>



<p>We can pass a function to each of these arguments, and&nbsp;<code>ggplot2</code>&nbsp;will use the value returned by that function for the corresponding aesthetic. If you pass a function to fun.data, you can compute many summary statistics and return them as a vector, where each element in the vector is named for the aesthetic it should be used for.</p>



<p>Let's understand this with two examples:</p>



<h3>Bar Chart with Median Values</h3>



<p>We will use the <code>stock_prices.tidy</code> dataframe we created earlier to plot a bar chart with the stock symbols on the x-axis and the median stock price for each stock on y-axis. We can achieve this using the <code>stat_summary()</code> function as follows:</p>



<pre class="wp-block-code"><code>ggplot(stock_prices.tidy,aes(x=Symbol,y=Prices,fill=Symbol))+
  stat_summary(fun.y = median, geom = "bar")</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-summary1.png" alt="" class="wp-image-27378"/></figure></div>



<h3>Quartile Points</h3>



<p>Following is another example where we plot quartile points for each stock. We first create a new function to calculate the quartile and then supply that function as argument to <code>fun.data</code> in <code>stat_summary()</code>.</p>



<pre class="wp-block-code"><code>median.quartile &lt;- function(x){
  out &lt;- quantile(x, probs = c(0.25,0.5,0.75))
  names(out) &lt;- c("ymin","y","ymax")
  return(out)
}
ggplot(stock_prices.tidy, aes(x=Symbol,y=Prices,col=Symbol)) +
  stat_summary(fun.data = median.quartile, geom = "pointrange")</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-summary2.png" alt="" class="wp-image-27379"/></figure></div>



<p></p>

stat_summary for Statistical Summary in ggplot2 R

<p>Faceting is a very useful visualization technique that allows you to create multiple versions of the same plot with multiple subsets of the data. This is also called small multiples. ggplot2 allows us to small multiples in two ways:&nbsp;<code>facet_wrap()</code>&nbsp;and&nbsp;<code>facet_grid()</code>. The default is&nbsp;<code>facet_null()</code>&nbsp;which produces a single plot.</p>



<p>We will use German Credit dataset to demonstrate faceting with both these functions.</p>



<h3>facet_wrap()</h3>



<p><code>facet_wrap()</code> creates a plot for every level of a factor passed to it. For example, if we create a scatter plot with Credit Amount on x-axis, Duration of Credit on y-axis, and pass the factor 'Type of Housing' as facet. Then it will create the number of plots equal to the levels in type of Housing with each plot displaying the points only for that level.</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount))
g+geom_point()+geom_smooth()+
  facet_wrap(~Type.of.housing)</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-facet-1.png" alt="" class="wp-image-27384"/></figure></div>



<p>You can control how the panels are wrapped into a grid with <code>ncol</code>, <code>nrow</code>, <code>as.table</code> and <code>dir</code>. <code>ncol</code> and <code>nrow</code> control how many columns and rows (you only need to set one). <code>as.table</code> controls whether the facets are laid out like a table (TRUE), with highest values at the bottom-right, or a plot (FALSE), with the highest values at the top-right. <code>dir</code> controls the direction of wrap: horizontal or vertical.</p>



<p>In the following example, we pass the factor 'Job' for faceting (which as 4 levels). We pass the parameter <code>nrow = 2</code> to wrap the plots in 2 rows.</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount))
g+geom_point()+geom_smooth()+
  facet_wrap(~Job, nrow = 2)</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-facet-2.png" alt="" class="wp-image-27385"/></figure></div>



<p>There are two important observations here. One, facets create subsets of the same dataset. Two, the x and y scales of each plot are the same in each facet.</p>



<h3>facet_grid()</h3>



<p><code>facet_grid()</code>&nbsp;allows use to facet using two variables.</p>



<h3><code>. ~ a</code> spreads the values of a across the columns as shown below:</h3>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount))
g+geom_point()+geom_smooth()+
  facet_grid(. ~ Loan.Quality)</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-facet-3.png" alt="" class="wp-image-27386"/></figure></div>



<h3><code>b ~ .</code> spreads the values of b down the rows.</h3>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount))
g+geom_point()+geom_smooth()+
  facet_grid(Loan.Quality ~ .)</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-facet-4.png" alt="" class="wp-image-27387"/></figure></div>



<h3><code>a ~ b</code>&nbsp;spreads a across columns and b down rows.</h3>



<p>It is recommended to put the variable with the greatest number of levels in the columns, to take advantage of the aspect ratio of your screen.</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount))
g+geom_point()+geom_smooth()+
  facet_grid(Type.of.housing ~ Loan.Quality)</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-facet-5.png" alt="" class="wp-image-27388"/></figure></div>



<h4>Facet Scales</h4>



<p>For both the facet functions, you can control whether you want the x and y position scales for each panel to be fixed across all panels (fixed) or allow to vary between panels (free).</p>



<ul><li><code>scales = "fixed"</code>: x and y scales are fixed across all panels.</li><li><code>scales = "free_x"</code>: the x scale is free, and the y scale is fixed.</li><li><code>scales = "free_y"</code>: the y scale is free, and the x scale is fixed.</li><li><code>scales = "free"</code>: x and y scales vary across panels.</li></ul>



<p>Usually you will want all facets to have same x and y scales for the purpose of comparability. However, in some cases free scales will make more sense, for example, in a situation where each level is measured on a different scale.</p>

Facets for ggplot2 Charts in R (Faceting Layer)

<p>The coordinate system of a plot refers to the space on which the data is plotted. The coordinate system of a plot, together with the x and y position scale, determines the location of geoms.</p>



<p>Below are the available coordinate options:</p>



<ul><li><code>coord_cartesian</code>: This is the default coordinate system (x horizontal from left to right, y vertical from bottom to top)</li><li><code>coord_flip</code>: Flipped cartesian coordinate system (x vertical from bottom to top, y horizontal from left to right)</li><li><code>coord_trans</code>: Used to to transform the coordinate system. We, i.e., substitute traditional axes with logarithmic axes and then present the values and statistics</li><li><code>coord_equal</code>: Ensures the units are equally scaled on the x-axis and on the y-axis</li><li><code>coord_polar</code>: Polar coordinate system; the x (or y) scale is mapped to the angle (theta)</li><li><code>coord_map</code>: Various map projections</li></ul>



<p>Most popular graphs such as line and bar charts are drawn using Cartesian coordinates.</p>



<h3>coord_cartesian</h3>



<p>When we plot a chart, Cartesian Coordinates is the default coordinate system. For example, the following chart plots Duration of Credit on x-axis and Credit Amount on y-axis. The points are colors by the factor Loan.Quality.</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount,color=Loan.Quality))
g+geom_point()+geom_smooth()+coord_cartesian()</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-coord-1.png" alt="" class="wp-image-27392"/></figure></div>



<h3>coord_flip</h3>



<p>By specifying coord_flip, it will still use the cartesian system, but flip the two axis.</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount,color=Loan.Quality))
g+geom_point()+geom_smooth()+coord_flip()</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-coord-2.png" alt="" class="wp-image-27393"/></figure></div>



<h3>Zooming In</h3>



<p>We can use the coordinates layer to zoom in to a plot by setting limits. The various options are below:</p>



<pre class="wp-block-code"><code>coord_cartesian(xlim = NULL, ylim = NULL, expand = TRUE)</code></pre>



<ul><li><code>xlim, ylim</code>: Limits for the x and y axes.</li><li><code>expand</code>: If TRUE, the default, adds a small expansion factor to the limits to ensure that data and axes don't overlap. If FALSE, limits are taken exactly from the data or xlim/ylim</li></ul>



<p>Let's say we want to zoom in our plot to show the points where the Duration of credit is between 40 and 60. This can be done as follows:</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount,color=Loan.Quality))
g+geom_point()+geom_smooth()+coord_cartesian(c(40,60))</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-coord-3.png" alt="" class="wp-image-27394"/></figure></div>



<h3>Aspect Ratio</h3>



<p>We can set the aspect ratio of a plot with <code>coord_fixed()</code> or <code>coord_equal()</code>. Both use <code>aspect = 1</code> (1:1) as a default. For example, consider the following plot which plots Duration of Credit on x-axis and Age on y-axis.</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Age.in.years,color=Loan.Quality))
g+geom_point()+geom_smooth()</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-coord-4.png" alt="" class="wp-image-27395"/></figure></div>



<p>We can add the coordinate coord_equal() to make the aspect ratio 1:1.</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Age.in.years,color=Loan.Quality))
g+geom_point()+geom_smooth()+coord_equal()</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-coord-5.png" alt="" class="wp-image-27396"/></figure></div>



<p></p>

Coordinates in ggplot2 in R

<p>The themes layer is used to style all the non-data ink of the plot, i.e., all the visual elements that are not part of data.</p>



<p>When we create a chart using ggplot2, it automatically uses a default theme. For theming we have a few choices:</p>



<ol><li>Use the default theme (What we have seen so far). The default theme is called&nbsp;<code>theme_grey()</code>.</li><li>Use a built-in theme. There are two built-in themes 1)&nbsp;<code>theme_grey()</code>&nbsp;which is default and 2)&nbsp;<code>theme_bw()</code>, a theme with a white background.</li><li>Modify/override specific elements of the default theme. This is done by using the&nbsp;<code>theme()</code>&nbsp;function. It overrides the graphical parameters of the default theme.</li><li>Create your own theme</li></ol>



<h3>Use the Default Theme</h3>



<p>The following chart is created using the default theme:</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount,color=Loan.Quality))
g+geom_point()+geom_smooth()+
  facet_wrap(~Type.of.housing)</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-theme-1.png" alt="ggplot2 chart with default theme" class="wp-image-27400"/></figure></div>



<p>Here is the definition of <code>theme_gray</code> theme, straight from ggplot2's source code:</p>



<pre class="wp-block-code"><code>theme_gray &lt;- function(base_size = 12) {
  structure(list(
    axis.line =         theme_blank(),
    axis.text.x =       theme_text(size = base_size * 0.8 , lineheight = 0.9, colour = "grey50", vjust = 1),
    axis.text.y =       theme_text(size = base_size * 0.8, lineheight = 0.9, colour = "grey50", hjust = 1),
    axis.ticks =        theme_segment(colour = "grey50"),
    axis.title.x =      theme_text(size = base_size, vjust = 0.5),
    axis.title.y =      theme_text(size = base_size, angle = 90, vjust = 0.5),
    axis.ticks.length = unit(0.15, "cm"),
    axis.ticks.margin = unit(0.1, "cm"),
    legend.background = theme_rect(colour="white"), 
    legend.key =        theme_rect(fill = "grey95", colour = "white"),
    legend.key.size =   unit(1.2, "lines"),
    legend.text =       theme_text(size = base_size * 0.8),
    legend.title =      theme_text(size = base_size * 0.8, face = "bold", hjust = 0),
    legend.position =   "right",
    panel.background =  theme_rect(fill = "grey90", colour = NA), 
    panel.border =      theme_blank(), 
    panel.grid.major =  theme_line(colour = "white"),
    panel.grid.minor =  theme_line(colour = "grey95", size = 0.25),
    panel.margin =      unit(0.25, "lines"),
    strip.background =  theme_rect(fill = "grey80", colour = NA), 
    strip.text.x =      theme_text(size = base_size * 0.8),
    strip.text.y =      theme_text(size = base_size * 0.8, angle = -90),
    plot.background =   theme_rect(colour = NA, fill = "white"),
    plot.title =        theme_text(size = base_size * 1.2),
    plot.margin =       unit(c(1, 1, 0.5, 0.5), "lines")
  ), class = "options")}</code></pre>



<h3>Use a built-in Theme</h3>



<p>We can use another built-in theme by adding the theme name to the plot function using a + sign.</p>



<p>The <code>theme_bw()</code> function produces a mostly (but not entirely) black-and-white theme.</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount,color=Loan.Quality))
g+geom_point()+geom_smooth()+facet_wrap(~Type.of.housing)+theme_bw()</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-theme-2.png" alt="ggplot - Chart with built-in black and white theme" class="wp-image-27401"/></figure></div>



<h3>Modify/Override specific elements of the chosen theme</h3>



<p>We can use the default theme or another built-in theme and then override some elements of it using the&nbsp;<code>theme()</code>&nbsp;function.</p>



<p>The visual elements can be classified as one of the three different types:</p>



<ul><li>Text</li><li>Line</li><li>Rectangle</li></ul>



<p>Each type can be modified by calling the appropriate function, namely,&nbsp;<code>element_text()</code>,&nbsp;<code>element_line()</code>&nbsp;and&nbsp;<code>element_rect()</code>.</p>



<p>Let's take a few examples to understand how this works.</p>



<h4>Change Plot Background to Light Yellow</h4>



<p>By default, the entire plot has a white background. We will change to to light green color. The background color of the plot is defined by <code>plot.background()</code> in the default function. The background is a part of the rectangle shape, so we can modify it using the <code>element_rect()</code> function. We will change the background color to light yellow and add a gray border of size 2.</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount,color=Loan.Quality))
plot&lt;-g+geom_point()+geom_smooth()+facet_wrap(~Type.of.housing)
plot+theme(plot.background = element_rect(fill = "#FFF6B0",color = "#737373", size = 2))</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-theme-3.png" alt="ggplot 2 - Chart Themes" class="wp-image-27402"/></figure></div>



<h4>Change Axis Line and Tick Color</h4>



<p>The plot lines can be modified using the <code>element_line()</code> function. Take a look at the default theme function above. The axis lines are <code>axis.line</code> and axis ticks are defined with <code>axis.ticks</code>. Also notice that the panels have a grid which is defined by <code>panel.grid</code>. We will remove the grids by using <code>element_blank()</code>.</p>



<pre class="wp-block-code"><code>plot+theme(plot.background = element_rect(fill = "#FFF6B0",color = "#737373", size = 2),
           panel.grid = element_blank(),
           axis.line = element_line(color = "#FF7E42"),
           axis.ticks = element_line(color = "#FF7E42")
           )</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-theme-4.png" alt="" class="wp-image-27403"/></figure></div>



<h4>Remove Panel and Legend Backgrounds</h4>



<p>We will now go a step further and remove the backgrounds of the panels and the legends as shown below:</p>



<pre class="wp-block-code"><code>g &lt;- ggplot(df,aes(x=Duration.of.Credit..in.months.,y=Credit.amount,color=Loan.Quality))
plot&lt;-g+geom_point()+geom_smooth()+facet_wrap(~Type.of.housing)
plot+theme(plot.background = element_rect(fill = "#FFF6B0",color = "#737373", size = 2),
           panel.grid = element_blank(),
           axis.line = element_line(color = "#FF7E42"),
           axis.ticks = element_line(color = "#FF7E42"),
           panel.background = element_blank(),
           legend.key = element_blank(),
           legend.background=element_blank(),
           strip.background = element_blank()
           )</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-theme-5.png" alt="" class="wp-image-27404"/></figure></div>



<h4>Change Text in the Chart</h4>



<p>We can make changes to the various text elements using the&nbsp;<code>element_text()</code>&nbsp;function. Let's make the following changes:</p>



<ul><li>Change strip text (text that appears in panels) to pink (#E1315B) and increase their size to 16.</li><li>Change color of x-axis and y-axis titles to blue (#008DCB) and make them italics</li></ul>



<pre class="wp-block-code"><code>plot+theme(plot.background = element_rect(fill = "#FFF6B0",color = "#737373", size = 2),
           panel.grid = element_blank(),
           axis.line = element_line(color = "#FF7E42"),
           axis.ticks = element_line(color = "#FF7E42"),
           panel.background = element_blank(),
           legend.key = element_blank(),
           legend.background=element_blank(),
           strip.background = element_blank(),
           strip.text = element_text(size = 16, color = "#E1315B"),
           axis.title.y = element_text(color = "#008DCB", hjust = 0, face = "italic"),
           axis.title.x = element_text(color = "#008DCB", hjust = 0, face = "italic"),
           axis.text = element_text(color = "black")
           )</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-theme-6.png" alt="" class="wp-image-27405"/></figure></div>



<h4>Change Legend Position</h4>



<p>We can modify various aspects of the legend using the <code>theme()</code> function. In the default theme, the legend has the following properties:</p>



<pre class="wp-block-code"><code>legend.background = theme_rect(colour="white"), 
    legend.key =        theme_rect(fill = "grey95", colour = "white"),
    legend.key.size =   unit(1.2, "lines"),
    legend.text =       theme_text(size = base_size * 0.8),
    legend.title =      theme_text(size = base_size * 0.8, face = "bold", hjust = 0),
    legend.position =   "right"</code></pre>



<p>As you can see, the default position is right. The other options are 'bottom', 'left', 'top' or 'none'.</p>



<p>In the following example, we change the legend position to bottom.</p>



<pre class="wp-block-code"><code>plot+theme(plot.background = element_rect(fill = "#FFF6B0",color = "#737373", size = 2),
           panel.grid = element_blank(),
           axis.line = element_line(color = "#FF7E42"),
           axis.ticks = element_line(color = "#FF7E42"),
           panel.background = element_blank(),
           legend.key = element_blank(),
           legend.background=element_blank(),
           strip.background = element_blank(),
           strip.text = element_text(size = 16, color = "#E1315B"),
           axis.title.y = element_text(color = "#008DCB", hjust = 0, face = "italic"),
           axis.title.x = element_text(color = "#008DCB", hjust = 0, face = "italic"),
           axis.text = element_text(color = "black"),
           legend.position = "bottom"
           )</code></pre>



<div class="wp-block-image"><figure class="aligncenter size-large"><img src="https://financetrain.com/wp-content/uploads/ggplot-theme-7.png" alt="" class="wp-image-27406"/></figure></div>



<p>This way we can modify almost every element of the ggplot and achieve any kind of styling we want.</p>

Changing Themes (Look and Feel) in ggplot2 in R

Data Visualization with R

This frontmatter description will appear below the title

Data Visualization with R

Learn how to create beautiful data visualizations in R using Base R graphics and ggplot2

What's Included

Course Resources

Lessons

Free Guides - Getting Started with R and Python

Accelerate your finance career with cutting-edge data skills.