Twitter social bots: The 2019 Spanish general election data

Area of Science:

• Computational Social Science
• Political Science
• Data Science

Background:

• Social bots are automated accounts influencing public opinion on social media platforms.
• Understanding bot behavior is crucial for analyzing political discourse and election integrity.
• Existing datasets may lack the depth or specific features needed for comprehensive bot analysis.

Purpose of the Study:

• To present a comprehensive, anonymized Twitter dataset from the 2019 Spanish general election.
• To facilitate research on social bot detection, analysis, and classification.
• To support the development and testing of machine learning models for identifying and understanding automated political influence.

Main Methods:

• Collected 5.8 million tweets from nearly 800,000 users discussing politics, identified via 46 hashtags.
• Anonymized user data to ensure privacy.
• Enriched the dataset with features including topic mentions, keywords (political bag-of-words), sentiment scores, bot likelihood, political affinity, and follower/following lists.

Main Results:

• A large-scale, feature-rich dataset specifically tailored for social bot research in a political context.
• The dataset includes user-level features indicating bot probability and political alignment.
• Tweets are annotated with topic information and sentiment, enabling nuanced analysis.

Conclusions:

• The dataset provides a valuable resource for researchers studying automated influence in political campaigns.
• Enables advancements in machine learning techniques for social bot detection and characterization.
• Facilitates a deeper understanding of the role of social bots in shaping public opinion during elections.