Measuring slums from space
Informal settlements are home to millions of people, yet they are often poorly mapped, making effective service delivery and policy planning more difficult. By combining satellite imagery with AI, researchers can identify and map these settlements accurately and at relatively low cost. This policy brief shows that models trained using data from cities in India and in Burkina Faso also perform well in other locations, reducing the need for costly ground surveys while supporting more timely policy decisions. However, local mapping remains essential for training, validating and improving these models.
Project
-
Gechter-Sifakis-Swanson-Tsivanidis-Policy-Brief-September-2025.pdf
PDF document • 327.26 KB
- Even though informal settlements house millions, they are poorly mapped, making service delivery and policy planning difficult.
- Satellite imagery combined with AI models can reliably identify informal settlements at low cost.
- Models trained in one city in India and in Burkina Faso also produce accurate maps in other cities, illustrating the potential to save on expensive ground surveys and deliver up-to-date policy responses.
- Local mapping efforts remain important for model training and validation.