HumAID: Human-Annotated Disaster Incidents Data from Twitter with Deep Learning Benchmarks

Text | General | Response

HumAID is a general-disaster dataset for categratization of humanitarian tasks post disaster. It contains 77,196 annotated tweets and is introduced for multiclass classification.

  • ML task type: Multiclass classification
  • Data Source: Social Media (Twitter)
  • Size: 77,196 Tweets
  • Timespan: 2016 - 2019
  • Geographical Coverage: Global
  • Baseline Information
  • Evaluated on: RF, SVM, FastText, BERT, D-BERT, RoBERTa, XLM-R
  • Metrics used: Weighted F1
  • Results as reported in original paper: Weighted F1: 0.781(RoBERTa)

Firoj Alam, Umair Qazi, Muhammad Imran, and Ferda Ofli. HumAID: Human-annotated disaster incidents data from twitter with deep learning benchmarks. In International Conference on Web and Social Media, pages 933–942, 2021.