Exploring and visualizing bias in policing Illinois ; initial data based on Stanford's Open Policing; we are also looking at exploring/analyzing Invisible Institute's Chicago Citizen's Police Data, but, at first glance, this is somewhat more segmented (and will need to be unified?)
Stanford's Open Policing (SOP) data from IL is from 2004-2015--a HUGE, very complete dataset.
It is available as a .tar.gz here or as a .zip here (NU Box folder/file).
There is also Census data and Polic Public Contact Survey data.
(A) Examining what factors affect if a driver in Illinois has their car searched for contraband
(B) Build some nice visualizations of the data to make some statements about potential disparities in traffic policing in IL + Chicago; perhaps push to Open Data Bits?
For (A), some preliminary analysis has already been done in an iPython notebook
For (B) there are at least two initial thrusts to analysis:
(1) 'Port' the R code used in the working paper from Stanford to Python
(2) Filter the data to focus on IL or Chicago (or Evanston)
Ideally, (1), could be done separately from (2), and then the methods from (1) can be reapplied to (2).
Logically, we will try to mimic the the file structure from Stanford's github for the national project
A demo for different visualization tools with DataFrame.plot APIs and Voyager has been added in /dotplot and /voyager, respectively