I do not pretend to be an expert on criminal or social justice. But as a topic that has been prominent in recent news and public opinion, I was curious about my local area's arrest statistics. Polk County, Iowa is predominantly white, and I wondered how arrest rates and other statistics varied by race, gender, etc.
If we assume that race plays no role whatsoever in a person's likelihood of being jailed, then we would expect the inmates' racial makeup to be pretty close to the county's overall demographics, which are as follows:
| White alone | 86% |
| Black or African American alone | 7% |
| American Indian and Alaska Native alone | 0% |
| Asian alone | 4% |
| Native Hawaiian and Other Pacific Islander alone | 0% |
| Two or More Races | 2% |
| Hispanic or Latino | 8% |
| White alone - not Hispanic or Latino | 79% |
So how do our arrest records look? I searched to see whether this information was already published, and found the following:
- The Des Moines Register - reports arrests from the last 60 days with some good summary reporting and visuals, but no reporting by race.
- Polk County's website has detailed records for each current inmate, but very thin summary reports.
Polk County's page was the most detailed, readily available source I found. However, the details are buried in each individual inmate's page, which would be a pain to gather by hand. Luckily, I know a thing or two about web scraping. I wrote a simple python script to crawl through Polk County's current inmate listing, gather data on each inmate, and perform analyses on the resulting data. The current inmate population's racial makeup was (as of 8/11):
| White | 68% |
| Black | 30% |
| Asian | 2% |
| Pacific Islander | <1% |
Some big differences there compared to the census data.
As the title image suggests, this is entry one in a series. Next time, I'll be slicing the data by more variables, and adding some visualizations. Here's a preview, with a plotly representation of the two tables above (hover over the bars to see values):