The Ethics of Data Visualization

Author: Peter Haferl

Publisher: Medium

Publication Year: 2019

Summary: The following article discusses how data science is not only about analyzing data and interpreting results, but also about telling a story around these results. Oftentimes, these stories include visualizations, that more clearly present the findings of an analysis. Haferl emphasizes the detail in which ethics are considered during the data analysis process, but that these considerations are less often made when thinking about presenting results. He outlines 2 main ways in which data scientists must be careful to act ethically during the visualization processes. The first one is to realistically portray findings without exaggerating or understating the result and the second is to not invert axis when portraying results visually to an audience. Cutting of the graphs at a convenient point to exaggerate a message, using incorrect quantities in visualizations, or changing aspect ratios to one’s personal benefit are all ways in which an audience can learn incorrect results about an analysis. All of these should be avoided when creating visualizations. The second part that Haferl discusses is the fact that data scientists should use the Cartesian coordinate system to enhance an audiences’ understanding of the results, and to avoid confusion around the interpretation of the results.