I am not from the USA, but I see a lot of discussions about race and racism in the USA and whether skin color is an indicator for crime, education, divorce and that sort of thing. I do not have political agenda (though I am sure that color in itself can hardly be a causality for anything), but I would like to make the multivariate analysis myself to get a better understanding about the topic. I do not trust the statistics (either way) from those guys in the talk-shows. Are there detailed datasets that I could analyses? I would prefer more fine-grained data, since I fear that aggregated data already is influenced by political agenda (either way). I am happy about data from any country, but I guess the USA is one of the more diverse countries on earth.
I've done some work in this area. The US FBI's data is aggregated from state reporting, and there are only certain types of crime they are required to report to the FBI (more info in this document https://ucr.fbi.gov/additional-ucr-publications/ucr_handbook.pdf/view).
Many US cities provide crime data via a public API. Here are some I've come across. Each city captures different information about the crime and the offender so try some out and see if it fits your needs.
- Philly: https://www.opendataphilly.org/dataset/crime-incidents
- Chicago: https://data.cityofchicago.org/Public-Safety/Crimes-2001-to-present/ijzp-q8t2
- SF: https://data.sfgov.org/Public-Safety/SFPD-Incidents-from-1-January-2003/tmnf-yvry
- Baton Rouge: https://data.brla.gov/Public-Safety/Baton-Rouge-Crime-Incidents/fabb-cnnu
- Providence: https://data.providenceri.gov/Public-Safety/Providence-Police-Department-Arrest-Log-Past-30-Da/vank-fyx9
- Atlanta: http://opendata.atlantapd.org/Crimedata/Default.aspx