TwitterDC

Author: Justin Cheng Copyright: 2010-2011

Synopsis

A set of ruby scripts designed to analyze graph data.

Requirements

Gems - ai4r

ForestLib

A set of classes which allow different graph-theoretic properties to be calculated, and also helper classes to parse source files.

TwitterDc

A set of Twitter-specific classes to specifically analyze Twitter data.

In main,

1. Run Atm1 to find the people who sent more than N messages to each other, and build the filtered graphs
2. Run Atm2 to see if more than k messages were sent both ways (reciprocated) or more than k messages sent one way but no messages the other way (unreciprocated)
3. Run Atm3 to build strongest connected component sizes for the unreciprocated graphs

Building the SCC Charts

1. Run Plot1 for the various stages i. Stage 1 calculates the # of unique nodes and depends on Atm2 ii. Stage 2 plots the SCC proportions and depends on Atm3

Prediction Graphs

1. Run Pred1 for the various stages in order i. Stage 1 calculates the degree count and builds the edge graph and depends on Atm1 ii. Stage 2 does prediction based on the generated reciprocated graphs and depends on Atm2 iii.Stage 3 generates predicted rec/unrec graphs iv. Stage 4 builds the SCC for the predicted unrec graph