Machine Learning for Peace:

Digital Tools for Civic Actors

DevLab@Penn

University of Pennsylvania

June 26, 2024

Principal Investigators: Jeremy Springman, Erik Wibbels

Data Scientists: Zung-Ru Lin, Hanling Su, Mahda Soltani, Jack Jacobs

Measuring Civic Space

MLP: Approach


How can data improve civic action?

  1. Awareness: data on what’s happening very recently
    • Mass scraping online news + ML to track events
    • Interactive data dashboard

  2. Planning: predictive analytics for strategic decisions
    • Forecasting political events
    • Civic Space Early Warning System

Data Production

Input: Online news

  • 400+ news sources
  • 40 languages
  • 120 million articles

Data quality

  • Focus on high-quality local sources
  • Much better coverage than extant archivers/aggregators (GDELT, Wayback, Lexis Nexis, etc.)



Output: Monthly data

  • 60 countries
  • 2012 - last month

Data Processing Pipeline



Measuring Civic Space Activity

Digital Tools

Quarterly Reports

Quarterly Reports

Detecting Events

Forecasting Events

Quarterly Reporting

Successful Early Warnings

  • ~70% success rate
  • ~80 events across 29 countries


Corruption:

  • Nigeria (Jan)
  • Angola (Dec)

Legal Actions:

  • Nepal (Feb)
  • Kyrgyzstan (Jan)


Legal Changes

  • Jamaica (Mar)

Non-lethal Violence:

  • India (Jan)

Purge/Reshuffle

  • Malawi (Jan)

Data Dashboard

Dissemination & Extensions

  • Distribution:
    • 3,500 unique visitors across 120 countries (annual)
    • ~150hr/mo active app usage time
    • Mailing list of ~600
  • Forthcoming Features:
    • Dramatically improved shock forecasts and new dashboard
    • Foreign influence event tracking and forecasts

Building on MLP’s Infrastructure

Extensions and Collaboration

  • Tracking new events and extracting new information
    • Environmental adaptation behaviors
    • Politically targeted anti-corruption campaigns
  • Policy research
    • Environmental adaptation and migration
    • Media capacity-building on article quality
  • Forecasting third-party event data
    • State Department Travel Warnings
    • Institute of Migration flow data

Influence Tools

Data Production

Influence Tools

Influence Tools

Extensions and Collaboration

  • Tracking new events and extracting new information
    • Environmental adaptation behaviors
    • Politically targeted anti-corruption campaigns
  • Policy research
    • Environmental adaptation and migration
    • Media capacity-building on article quality
  • Forecasting third-party event data
    • State Department Travel Warnings
    • Institute of Migration flow data

Appendix

Planning: Forecasting events

Underlying Data: Crawlers vs Direct

Scraping: Wayback vs Custom

El Diaro

mlpeace.org