Global Civil Unrest on Twitter (G-CUT): Geotagged Tweets from 42 Countries, 2014-2019

Geotagged tweets from the Twitter streaming API collected from 2014 to 2019 for 42 countries across Africa, the Middle East, and Southeast Asia. Only English tweets identified by langid are included. The dataset was created by Abhinav Chinta of Johns Hopkins University for the study of civil unrest manifestation on social media.

Use Cases

• Analyze temporal patterns of civil unrest discourse based on the 2014-2019 date range
• Map geographic hotspots of unrest-related conversation based on geotagged tweet locations
• Study English-language protest narratives from specific regions like Africa, the Middle East, and Southeast Asia
• Train models to detect civil unrest signals from social media text

Strengths

• Data spans a 6-year period from 2014 to 2019
• Covers 42 countries across three major world regions
• Tweets are geotagged, enabling spatial analysis
• Associated with a published academic paper and GitHub repository for details

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count is unknown, which may limit suitability assessment
• Limited to English tweets only, which may exclude local language discourse

Provenance

Source

Twitter streaming API

Collection Method

Collected via the Twitter streaming API, filtered for English language using langid.

Time Range

2014 to 2019 inclusive

Geography

42 countries from Africa, the Middle East, and Southeast Asia