Football Fan Communities and Political Influence on X (Twitter): An Anonymised Research Dataset (2016–2017)

This repository contains an anonymised research dataset supporting the study of political influence within online football fan communities. The dataset underpins the published paper "Kicking Politics: How Football Fan Communities Became Arenas for Political Influence" and the ongoing research project "Kicking Politics: Towards a Framework for Detecting Political Influence in Online Communities." The original dataset comprises approximately 152.3 million football-related tweets collected through the Twitter Streaming API between 25 July 2016 and 26 October 2017. Tweets were collected using accounts and hashtags associated with 44 clubs from the English Premier League and English Football League Championship. From this corpus, an iterative content-filtering procedure identified 94,846 politically relevant tweets for detailed analysis. These tweets were produced by 43,252 unique user accounts and include: 41,447 original tweets (44%) 33,018 retweets (35%) 13,381 quote tweets (14%) 7,000 replies (7%) The dataset also contains 15,930 unique hashtags. Frequently occurring political hashtags include brexit, ukip, trump, edl, nhs, bnp, maga, ge2017, ira, indyref2, putin, euref, snp, labour, and donaldtrump. The Zenodo archive includes: an anonymised tweet-level dataset containing tweet metadata, interaction types, and anonymised user identifiers; a Gephi project (.gephi) containing the interaction network with anonymised nodes, edges, node attributes, and network metadata. To protect user privacy, all user and tweet identifiers have been replaced with deterministic pseudonymous identifiers generated using a salted cryptographic hash. The same original identifier is consistently mapped across all files, preserving the relationships between tweets, users, and network structures while preventing recovery of the original identifiers. Public account labels are retained only where appropriate to support reproducibility. The dataset is intended to support reproducible research on online communities, political communication, network analysis, and information diffusion. It may also serve as a benchmark dataset for the development and evaluation of methods for detecting political influence in online social networks. Any reuse should comply with the applicable platform terms of service and relevant ethical and legal requirements. If you use this dataset, please cite this Zenodo record together with the associated publication(s).