This is a Django version of "radar," a site and API that I am continually tinkering with to provide crime statistics for locations in Portland, Oregon.

This version uses ElasticSearch, geohash grids and median averages to compare crime at the user's location to average crime across the city.

The comparison is done by aggregating city crime data by geohash cell, summing by crime type for each cell (a bounding box), finding the median values for all crime types across cells, and then later finding the cell the user's location is in and comparing crime sums in that cell to the city averages.

First, set up the Vagrant instance:

$ vagrant up

Then, log in and make sure ElasticSearch is running:

$ vagrant ssh
$ sudo service elasticsearch start

Set up the index and data (still assuming you're in the Vagrant machine):

$ source ~/radar/bin/activate
$ cd /vagrant
$ cd crime_stats
$ python create_index.py
$ python load_crimes.py ../data/crimes_2013.json

Next, set up the Django database:

$ cd /vagrant
$ ./manage.py syncdb
$ ./manage.py migrate

And get your static files collected:

$ ./manage.py collectstatic

Log into the Vagrant machine in and run the development server.

$ vagrant ssh
$ source ~/radar/bin/activate
$ cd /vagrant
$ ./manage.py runserver 0.0.0.0:8000

The Vagrant file sets a static IP for the machine, so browse to http://192.168.50.4:8000/ to see the site.