Data Science: The Hard Parts

Chapter 8. Datavis: Choosing the Right Plot to Deliver a Message

Chapter 7 went through some good practices to build and deliver powerful narratives in data science. Data visualizations (datavis) are powerful tools to enrich your narratives and are a field of study on their own. As such, they need to be chosen as communication devices. The question you should always ask yourself is: is this plot helping me convey the message I want to deliver? If the answer is negative, you should go back to the drawing board and find the right plot for your message. This chapter goes through some recommendations that will help you improve your visualization skills.

Some Useful and Not-So-Used Data Visualizations

The field of datavis has evolved quite a bit in the last few decades. You can find online references, catalogues, and taxonomies that should help you find the right type of graph for your question. You can check the Data Visualisation Catalogue or from Data to Viz.

Unfortunately, many practitioners stick to default alternatives such as line and bar plots, often used interchangeably. In this chapter, I’ll review some less well-known types of plots you can use, and discuss some pitfalls that are common among data practitioners. This is in no way exhaustive, so at the end of this chapter I’ll point to some great resources that will provide a more complete picture of the field.

Bar Versus Line Plots

Let’s start with the most basic question of all: when should you use bar and line plots? ...