Data visualization supports a different kind of understanding than looking at a page of raw data, especially for viewers who may not be subject experts. Visualizing can help you emphasize a point, allow viewers to explore data they couldn't in its raw form, or make surprising connections. Visualizing your data may help:
It may be the only thing someone remembers after a presentation. Additionally, visualizing your data as a step in data evaluation or analysis can help you see patterns that you don't get from summary statistics!
Visualization, then, is a highly valuable tool for anyone communicating or analyzing data. There are as many different types of graphs and plots as there are use cases for visualization.
Anscombe's Quartet and the Datasaurus
These are two famous examples of datasets having the same summary statistics but very different visualizations. Anscombe's Quartet was created to show the effects of outliers and the importance of plotting your data when analyzing it. The descriptive statistics for each dataset are the same, within 2 decimal places, yet the plots look completely different.
Image from Wikimedia Commons - CC BY SA
The Datasaurus dozen collection of datasets is a modern equivalent of Anscombe's quartet, created by Alberto Cairo in 2016. It extends and exaggerates the point made by Anscombe by first using a tool called DrawMyData to create a plot of a T-rex, and then providing 12 alternative datasets that again have the same descriptive statistics but wildly different plots.
These two collections of datasets and plots demonstrate the most obvious reason to visualize your data - by seeing the data plotted, you may notice patterns, outliers, or other important aspects that could be missed with descriptive statistics alone.
Storytelling with data
Your interest in visualization may be in service of communication, in scientific, business, or journalistic contexts. Visualization is a powerful tool for helping people grasp the point of the data you have, and considering the story you want to tell is one of the earliest steps in the data visualization design process (discussed on another page of this guide). There are resources for all of the previously-mentioned contexts and more available either openly on the web or through the Leatherby Libraries.
