[phyrrus9]

So, I recently found a source for raw data coming from my local police department's database (legal source, but not disclosing it). I originally wanted it so that I could map crime hotspots, and the result was the following webapp



To get that map I throw out a lot of data, including:

Welfare checks
suicide attempts
medical calls
fire
traffic
commercial alarms
a bunch of unknown types
anything to do with juveniles

I then realized, I have traffic stop data, and I can look at just that data over a period of 30 days and map out hotspots where the cops patrol. The result was the following (made by hand) map



Notes: each point represents a specific coordinate where 3 or more stops were made in the last month. shaded areas are where those points all come within 1,000ft of each other

Let me know what you all think about the idea.

[phyrrus9]

Update: I've completely automated everything now, and shoved it onto a quick webserver. Page updates daily at a random time. https://pdmap.ipseitysoftware.com/

[Drako]

If you live on Montana St, you might as well be living in the slums. Anyways, that's a very neat way of using police data. What language did you use to represent the data?

[phyrrus9]

If you live on Montana St, you might as well be living in the slums. Anyways, that's a very neat way of using police data. What language did you use to represent the data?

The data comes in direct from PD as a very messy CSV file with a lot more data than just traffic. I'ts parsed out, filtered, and then shoved into a T-SQL database using a backend worker in C#.
Once there, it's geocoded (the info given has no coordinates and often bad addresses), and mapped to hotspots using straight queries.
After that, the worker computes polygons of hotzones and puts those back into the database.

The web frontend is a simple C# linux application that reads out the points and polygons and just outputs JS for the google maps API to take over.