twitter-cluster

Clustering twitter conversations

This repository contains code that facilitates looking at clusters of conversation within the twitters, and how they evolve over time.

Relevant files include:

• Conversational Cluster Detection and Transition Likelihood Calculation Python Version
• Conversational Cluster Detection and Transition Likelihood Calculation Unicage Version
• Performance Comparison between python and unicage
• Data collection scripts
• Visualization

config.json

To configure the locations of various data, working, and tools directories, modify the config.json.template file to include the relative paths to:

• data_dir : The location of raw twitter message dumps
• python_working_dir : Where intermediate python files should be stored
• unicage_working_dir : Where intermediate unicage working files should be stored
• cos-parallel : the path to the cos-parallel tool
• maximal_cliques : the path to cos-parallel's utility 'maximal cliques'

as an example:

{"maximal_cliques": "tools/cosparallel-0.99/extras/./maximal_cliques", 
 "cos-parallel": "tools/cosparallel-0.99/extras/./maximal_cliques",
 "python_working_dir": "../PYTHON/working/",
 "unicage_working_dir": "../UNICAGE/working/",
 "data_dir": "../data/" }