You work for a management consulting firm as a machine learning specialist. You are on a team of data scientists and other machine learning specialists. Your team has been assigned the task of building a machine learning model to predict Return On Investment (ROI) for new potential engagements that your management consults may wish to take onto their book of business.
You have a dataset of past engagements with many features that can help you define your problem as a machine learning problem. Before you decide on which machine learning algorithms to evaluate, you wish to visualize the historical data to get an idea of the relationships between three of the key features of your dataset: ROI, investment time, and investment size.
Which type of visualization would best give you an idea of the relationship between these three features?

Answer options:

A.Pie chart
B.Tree map
C.Column histogram
D.Bar chart
E.Bubble chart
F.Line chart

Answer: E
Option A is incorrect. A pie chart is best used to show the portion of the total for each slice of the pie. This type of chart doesn’t work well with three dimensions, such as ROI, investment time, and investment size.
Option B is incorrect. A tree map chart also shows the portion of the total. This type of chart is good for data with a long tail. But it also would not work well on three dimensions.
Option C is incorrect. Column histograms are distribution charts. They show how data is distributed at intervals. But you a looking for visualization to show the relationship between three variables.
Option D is incorrect. A bar chart is a comparison chart. These types of charts are good for showing how feature values change over time or to show a static snapshot of how different variables compare with each other. But you are looking for the relationship between three variables, not change over time or a static snapshot comparison.
Option E is correct. A bubble chart is a relationship chart. For a relationship between two variables, you could use a scatter chart. For the relationship between 3 variables, a bubble chart shows the relationships as such: x-axis for investment time, y-axis for ROI, and the bubble size for investment size.
Option F is incorrect. A line chart is used to show a comparison of variables changing over time. You are looking for a relationship between three variables, not how they change over time.
Reference:
Please see the AWS Data Visualization page, and the Amazon QuickSite overview page.