As a data scientist, I’m always interested in how statistical analysis can shed light on complex societal issues. Recently, I dug into the Washington Post’s database on fatal police shootings to better understand the data patterns behind this controversial subject for our project 2
The Post has done an impressive job compiling a comprehensive dataset based on public records, news reports, and original reporting. It contains over 6,000 fatalities since 2015, with dozens of attributes on each incident including victim demographics, whether they were armed, and contextual details. My analysis revealed alarming racial disparities in the data. While Black Americans account for less than 13% of the US population, they represent over 25% of those killed in the database. In contrast, fatality rates for White Americans closely align with their population share. This suggests systemic bias against people of color. Examining the subset of unarmed victims provided further evidence of risk disparity. Despite making up only 6% of the US population, Black Americans accounted for nearly 35% of unarmed civilians shot and killed. This implies non-violent Black citizens bear significantly higher risk of being killed by police. Additionally, my time series analysis showed fatal shootings have stayed relatively steady nationwide since 2015, averaging nearly 1,000 per year. Breaking this down by race, White deaths have risen slightly over this period while Black deaths have fallen but continue showing vast overrepresentation.
In summary, by letting data speak for itself, this project provides quantified insights into racial gaps in deadly police encounters. My assessment is that comprehensive reform is required to address these disparities and biases. But progress will only occur once the data-driven reality of the problem is acknowledged.