Mining the Geo Needles In The Social Haystack

Geo data is more than ever a staple of the social web and is exposed by popular web properties such as Facebook, Twitter, and LinkedIn. The aim of this workshop is to give attendees a very quick orientation and then get right business harvesting, store, analyzing, and visualizing geo data.

The overall goal of this presentation is to be very hands-on and empower attendees to get down and dirty with their geo data and answer questions such as the ones described in the proposal. About 2/3 of the workshop involves getting oriented to the space and slurping down data into a consistent JSON-based storage schema. The last 1/3 of the workshop introduces basic analysis and visualization ranging from plotting data on an online map to performing data mining techniques such as k-means and hierarchical clustering on the data.

Data munging is completed in Python; storage is managed by pragmatic storage technologies such as CouchDB and/or Redis; visualization is accomplished with popular mapping technologies such as Google Maps and Google Earth.

This workshop is loosely adapted from “LinkedIn: Clustering Your Social Network for Fun (and Profit?!?)” from Mining the Social Web - http://oreilly.com/catalog/0636920010203 - but is tailored specifically for Where 2.0 with a broader scope.