controversy-detection

Code and networks used in the paper "Quantifying Controversy in Social media" Author: Kiran Garimella (kiran.garimella@aalto.fi)

Structure

code/ # contains all the code controversy_measures

a. Code for computing polarization scores:

1. edge_betweenness/ 
	computePolarizationScoreEdgebetweenness.py
2. randomwalk/
	computePolarizationScoreRandomwalk.py
3. force_directed/
	computePolarizationScoreForceDirected.py
	forceatlas.py -- Code for computing the Force directed embeddings
4. GMCK/
	computePolarizationScoreICWSM.py
5. MBLB/
	computePolarizationScoreVenezuela.py

All these scripts for computing polarization scores depend on the presence of a "communities_{retweet,follow}_networks" folder. See below.

b. Additional folders:

1. community_detection_retweet_networks/ -- contains communities obtained from the retweet networks in ../networks/retweet_networks
2. community_detection_follow_networks/ -- contains communities obtained from the retweet networks in ../networks/follow_networks
3. edge_betweenness/ -- edge betweenness of the networks, used by the edge_betweenness polarization measure.

Community detection was done using Metis. Any other community detection can be used to obtain 2 communities.

---

networks/ # contains all the netwroks retweet_networks/ # all retweet networks we created (only the largest connected component) follow_networks/ # all follow networks we created