After a brief chat on Facebook regarding the recent spike of shootings in Chicago which resulted in 91 deaths in August 2016 I have promised to look at relevant data to see which claims can or cannot be backed up.
After aggregating some data from the City of Chicago Data Portal:
Crimes - 2001 to present, 1.5GB exported dataset using the chigufi.go
Go language script the resulting data and charts are summed up in the Google spreadsheet:
The sheets in the document show:
- How the IUCR (Illinois Uniform Crime Reporting) codes were grouped for the purposed of the aggregation. IUCR codes not listed in the sheet are not used in the aggregates.
- Graphs and the whole dataset aggregation from January 2011.
- Graphs of the aggregation of the recent data subset from January 2013.
- Terms of use and data disclaimer as required by the City of Chicago
Some details are off and I have no time to address them - for example, a number of homicides resulting from shootings may be higher than a reported number of murders "on the spot" in the police reports. I do not know if police retroactively repopulates these reports with newly uncovered information, and such.
I still think these charts are useful to provide context to the bare publicized number of homicides - even considering the shortcomings of this a-few-hours-hobby script.
There is intentionally no commentary or what to make out of these numbers. Interpreting the results requires a basic understanding of what is a relationship between correlation and causation, or what line regressions are.
To complete the references, here is the new Facebook post linking to the data.