1 of 9

Social Networks

Datasets and proposals

2 of 9

Twitter data

Main projects: �1 Dataset for Retweet & Mention networks

1 Dataset for User – News bipartite networks

3 of 9

Retweet & Mention network

  • Argentinian dataset from an online debate in Twitter during 2019.
  • Context:
    • Vicentin, a major Argentine soy exporter, defaulted a huge debt in 2019 and start bankruptcy proceedings. There was a political dispute about whether the state should nationalize the company or let it go bankrupt.
  • Data: Twitter dataset from 2019
    • # of users: 75.169 (14.664)
    • # of tweets and retweets: 205.214 and 171.962 (20mil and 16.758)
    • # of mentions: 11.205
  • Proposal: Analyze the retweet and the mention networks extracted from the data
  • Extra data:
    • https://networkrepository.com/rt.php

4 of 9

Guiding Questions

  • Create and plot the retweet and the mention networks
  • Characterize both of them using different network metrics
  • Analyze the community structure of both networks
  • Implement different kind of filters in order to analyze different network structures
  • Identify and characterize clusters
  • Identify key role players in the observed structure
  • Some ideas from:

Cita: Conover, M., Ratkiewicz, J., Francisco, M., Goncalves, B., Menczer, F., & Flammini, A. (2011).

Political Polarization on TwitterProceedings of the International AAAI Conference on Web and Social Media5(1), 89-96.

https://doi.org/10.1609/icwsm.v5i1.14126

5 of 9

The users-news network

  • Argentinian dataset from political active users in the period previous to national presidential elections in 2019
  • Data: Twitter dataset from 2019
    • # of users: 6.781
    • # of tweets: 73.640
  • Proposal: Analyze the bipartite users-news network extracted from the data
  • This data was used in:
    • “News-sharing on Twitter reveals emergent fragmentation of media agenda and persistent polarization”. Tomas Cicchini, Sofia Morena del Pozo, Enzo Tagliazucchi y Pablo Balenzuela. EPJ Data Science 11, 48 (2022). https://doi.org/10.1140/epjds/s13688-022-00360-8
  • Extra data:
    • https://networkrepository.com/rt.php

6 of 9

Guiding Questions

  • Provide relevant metrics characterizing the dataset
  • Build the bipartite user-news network
  • Build the bipartite user-media network
  • Calculate the proyections on user and news layer
  • Analyze the community structure of both networks
  • Identify and characterize clusters
  • Identify key role players in the observed structure

7 of 9

YoutTube data

Project: The video networks co-commented by users

8 of 9

Data from main US media outlets

  • Dataset with videos from 18 US media outlets (5 conservative, 4 neutrals and 9 liberals)
  • 3224 videos with comments and metrics (# of likes, # of views, # of comments, duration)
  • Proposal: Analyze the co-commented video network:
    • Nodes: videos
    • Edges: # of users co-commented two videos

9 of 9

Guiding Questions

  • Characterize the datasets through different metrics
  • Build the video-comented network
  • Characterize both of them using different network metrics
  • Analyze the community structure of both networks
  • Identify and characterize clusters
  • Identify key role players in the observed structure