Researchers Use Social Media To Predict Displacement
A University of Notre Dame study published in EPJ Data Science in 2024 analyzes nearly 2 million X posts across Ukraine, Sudan and Venezuela to predict forced displacement timing and volume using NLP. It finds sentiment signals and pretrained language models outperform emotion-based approaches, especially for conflict-driven cross-border movements, suggesting these tools can serve as early warnings when combined with traditional humanitarian data.
Key Points
- 1Analyzed nearly 2 million X posts across Ukraine, Sudan, and Venezuela to predict displacement.
- 2Found sentiment outperformed emotion for timing and volume prediction, especially for cross-border movements.
- 3Recommend pretrained language models as effective early-warning tools complemented by traditional data sources.
Scoring Rationale
Applies peer-reviewed NLP methods to humanitarian early warning, but limited by language coverage and variable crisis contexts.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
