Global Tech UpliftLinking Intelligence
Public Health Monitoring
Led at Boston Children's Hospital. Built and deployed an NLP pipeline that scans Twitter for food-poisoning signals, engages affected users, and routes verified clusters to health-department dashboards.
Continuous ingestion of tweets mentioning food-poisoning symptoms or related keywords, filtered by location, via the Twitter API.
ML models process tweet text to distinguish genuine foodborne-illness reports from unrelated content.
The system geocodes tweet origins to map incidents to specific health-department jurisdictions.
Automated outreach contacts tweet authors, encouraging formal illness reports to authorities.
Flagged cases are compiled and clustered to identify outbreak patterns.
A dashboard displays mapped incidents with case details and inspection status for health officials.
Officials use the alerts to prioritize restaurant inspections and outbreak investigations.
Cut outbreak detection and response time by about 60% across U.S. and U.K. health departments; documented in three peer-reviewed papers (2018-2020).