saulshanabrook
commented
6 years ago I started this analysis in the https://github.com/thelabdc/lights-and-crime/blob/master/Days%20before%20and%20after.ipynb notebook, but am still using repairs.xlsx and havent looked at iSlims yet.
The main goal of this first analysis is to count the total number of crime near a streetlight for each day before and after a repair. Saul has worked out the basics with public data in the first_analyses notebook using the "repairs.xlsx" data from DDOT, but his code aggregates all the days for a particular range. I think a more useful analysis would have a daily count for every day before and after the repair up to, say 30 days. You can use the repairs.xlsx data if you like, but I'd prefer you give the iSlims data a try, as it is closer to what we'll end up using. The only issue there is how to convert the X and Y to the right projection, and it's not clear to me what projection it's currently in. It would also be nice to have separate counts for crimes committed during the day, during the night, and maybe even at dusk. If you would create a new "crimes-per-day.ipynb" notebook to that end and give it a whirl, that would be lovely. Anyways, have fun and let me know if you have any questions. :-)