Machine learning for blue line tracing and wastewater surveillance

Project Lead: Andy Tatem

Team: Shengjie Lai, Somnath Chaudhuri, Alex Cunningham, Kathryn Baxter
Funding: University of Virginia
Start date: April 2025

Completion: December 2026

We are working with leading international research groups to create smarter, scalable ways to track infectious diseases in low and middle-income countries, using environmental data. By building on five years of multi-pathogen surveillance data from Dhaka, Bangladesh (2019–2024), we are identifying the bare minimum data needed to run effective environmental surveillance systems in places with limited resources.

The project will roll out in phases, starting with the development and testing of low-cost, high-impact tools. These include AI and machine learning techniques to map sewage networks – pulling detailed features from satellite and drone images to pinpoint the key infrastructure needed for tracking pathogens in wastewater.

Our team will take the lead on modelling small-area urban populations, using Dhaka’s rich, ground-level population data as the benchmark. The goal is to find the smartest ways to gather just enough on-the-ground data and combine it with geospatial tools to accurately estimate urban populations – without the heavy cost and time of full enumeration.

Once we’ve tested and proven these methods in Dhaka, we plan to apply them in selected African cities. Pending future funding, we aim roll out the optimized tools and strategies in cities where environmental surveillance efforts are already in motion – potentially in Nigeria, the Democratic Republic of Congo, and South Africa.

Project outputs

  • 1×1 km and administrative level risk maps of GWD for the six study countries and appropriate visualizations.
  • Threshold exceedance probability maps to identify epidemiological hotspots at the grid and administrative levels and appropriate visualizations.
  • Integration of the modelled risk maps with gridded population data to quantify populations living in hotspot or high-risk areas and recommendations for targeting interventions.
  • Results of analyses exploring the impact of interventions, population mobility and climate change on GWD distribution.

Partners

Link image

Sanitation and hygiene in Rohingya camp by Rabiul Hasan/icddr,b – CC BY-NC-ND 2.0