For last week’s Data Diary, Knarr CTO Speros Kokenes tackled some data around a very important issue, police use of force in the United States. After several weeks of protests following some highly publicized deaths of black citizens at the hands of police, many people are discussing racism and the role of race in police use of force, especially fatal force. To explore this issue, Speros used Knarr to review The Washington Post database of police use of force which the publication has been keeping on an ongoing basis since 2015. Speros pulled that data into Knarr as of June 2, 2020.
Speros first examined the question “what does it mean to be armed”. To begin, he created the chart of Fatal Shootings by Weapon Type [the weapons that “armed” individuals were carrying when killed by police]. At the top of the list are well-known violent instruments such as guns/knives, but if you continue down you see things like “toy weapons”– which themselves account for 184 deaths of “armed” people by police. You can also find things that seem ordinarily non-threatening like a screwdriver or pen. These raise the question: Was fatal force necessary in these instances?
Next, Speros looked at demographic differences between armed and unarmed shootings. Using Knarr, he separated incidents into the categories of armed vs. unarmed (with the knowledge in mind that what is considered “armed” by these classifications is questionable at times). Then Speros created a mixed analysis of the percentages of people that were armed and killed vs. people that were unarmed and killed including demographic breakdowns by race.
You would expect the demographic breakdown of fatal shootings of unarmed or armed people to be the same, but as you can see in this analysis this is not the case. In the breakdown of armed people killed by police 23% of victims are black and 46% are white, but in the breakdown of unarmed people killed by police, the percentage of black people jumps by 12 points up to 35%.
This raises more questions about why this difference is there, what causes it, and whether further exploration of the data might yield more insights around these questions. Other notable points a person could explore in this data set include the rate of people killed while fleeing vs. attacking.
As with any issue, we hope that by examining the data we are able to shed more light on the situation and dig into the truth of the matter. With our vision for Knarr, we hope that people are able to find insights that will help them in their pursuit of greater knowledge and making better decisions for the future, whether those are decisions related to public policy, business, or anything else.
This data is available at The Washington Post website, and The Washington Post also keeps a GitHub repository where they are storing data related to these incidents. You can also view Speros’s published analysis and commentary in the Knarr Note Viewer here.