Separating Polarization from Noise: Comparison and Normalization of Structural Polarization Measures
The data and scripts for Separating Polarization from Noise: Comparison and Normalization of Structural Polarization Measures-paper.
The network_data contains two subfolders (single_hashtag_networks and multiple_hashtag_networks). All the networks are in .edgelist format. The folder contains 183 topic endorsement networks inferred from Twitter interactions during the 2019 Finnish Elections. Nodes are accounts and undirected ties indicate uni- or bi-directional endorsement via retweets on the given topic. Please see [1] and [2] for details. No identifying information nor original raw data from the Twitter platform is included here. See the network_info.csv for more details on each network.
Please, find the properly anonymized and documented dataset on Zenodo (https://zenodo.org/records/8434372).
polarization_algorithms.py: the implementations of all polarization measures analyzed in the paper
improved_polarization_algorithms.py: faster and better documented implementation of the polarization algorithms [Recommended]
partition_algorithms.py: the different partition algorithms used for obtaining the two groups
dk_analysis.py: computing the polarization scores for randomized networks
deghet_analysis.py: the analysis of non-homogeneous degree sequences on polarization scores
[1] Chen, Ted Hsuan Yun ; Salloum, Ali ; Gronow, Antti ; Ylä-Anttila, Tuomas ; Kivelä, Mikko. / Polarization of climate politics results from partisan sorting : Evidence from Finnish Twittersphere. In: Global Environmental Change. 2021 ; Vol. 71. (https://doi.org/10.1016/j.gloenvcha.2021.102348)
[2] Salloum, Ali ; Chen, Ted Hsuan Yun ; Kivelä, Mikko. / Separating Polarization from Noise: Comparison and Normalization of Structural Polarization Measures. In: Proc. ACM Hum.-Comput. Interact. 6, CSCW1 ; 2022 ; Article 115 ; Vol 6. (https://doi.org/10.1145/3512962)
Please, cite the first paper if you use the network data, and the second one if you use the normalized polarization scores. Thanks!