This is an attempt at locating "hot spots" around you. Each pin represent a user action based on a particular location around San Francisco. We push pins to a web service running on tornado, and display these pins on a Google Map. The pins pushed to tornado aren't persisted, since the data can be recreated by using ElasticSearch. There are 2 files in this project: > main.py: - Generates as many locations as asked, and feeds then to the web service > hotspot.py - A web service running with tornado, ready to absorb every single pin you feed it, and display it on a map. It tries to be smart enough to not repin a hotspot too close to another hot spot. Runs by default on port 8888. # How to get started ? You need to have ElasticSearch running. Details about how to do so : http://www.elasticsearch.org/guide/reference/setup/installation.html 1. Create a new index: > curl -XPUT http://localhost:9200/pins 2. Generate the associated mapping > curl -XPUT 'http://localhost:9200/pins/pin/_mapping' -d ' { "pin" : { "properties" : { "location" : { "type" : "geo_point" } } } }' 3. Start the web service > ./hotspot.py --prod=False --logging=debug 4. Feed data to ElasticSearch, and pass it on to the Web Service > ./main.py --config=config.yaml --count=100 5. Open your browser and watch! http://localhost:8888 Here we can see that even with a lot of actions all over the map, we still manage to find out the 2 main "hot spots" of San Francisco http://dl.dropbox.com/u/12554984/7d34.png As we zoom in, it becomes obvious that something is happening at that particular location : From: http://dl.dropbox.com/u/12554984/40fb.png To : http://dl.dropbox.com/u/12554984/f98c.png Same goes for the AT&T park : http://dl.dropbox.com/u/12554984/16bf.png That's it for now.