The following project was designed as part of his thesis for the Aristotle University of Thessaloniki in the Department of Informatics. We present an analysis for finding echo chambers in widespread social media, using state-of-the-art datasets, which are available for use as created by the author for this purpose.
In the src folder you can find:
- analize_communities.py: presenting the partition in smaller communities.
- analyze_baltimore_riots.py: presenting results of baltimore_riots community.
- analyze_vacc_final.py: presenting results of vacc_final community.
- compute_communities.py: calculated communites and colored them in the Cytoscape files.
- compute_communities_hierarchical.py: created communities using the multilevel Louvain algorithm.
- pull_data_from_db.py: parse data from mongodb cluster
- wordcloud_communities: added wordclouds of all words in tweets
In the folder of this repo you can find,
communities/ : nodes and edges from all the networks
communities_hierachical/ : multilevel Louvain algorithm partitioning
cytoscape_files/ : visualization of the grapg using Cytoscape tool
echo chambers/ : final analysis after re-running the random walks algorithms
figures/: final figures
networks: final networks
new_worldlouds/: worldclouds that emerged after analysis
sentiment/ : polarity and subjectivity of the tweets
and launch the following command:
python main