SEEDS · Gramener

The problem

Warnings and other risk-related information are often vague and not up to date, covering macro-level areas and hard to understand by at-risk populations.

What was deployed

Vendor

Gramener

Products

Sunny Lives AI model, Kepler, Microsoft Azure

Technique

Classical ML

Build or buy

Built in-house or to order

Deployment

Custom build

Data used

high-resolution satellite imagery, manually tagged over 50,000 houses, geographic layers (Waterbodies, Distances from Road Network, Topographic Wetness Index (TWI), Elevation and Slope Vegetation (NDVI), Impervious surface, Landslide Risk, Building Footprints), historical data, open datasets

Scale

50K individuals from at-risk communities

Timeline

4 months

What changed

Each row is quoted from the source. Figures we could not find on the page in those words are marked — they are kept, not deleted, so you can judge them.

Difficulties and limits

• The population in those areas was dense enough to make accurate differentiation between houses difficult.
• As geographies change, so do the roof types, and identifying these different roof types was of prime importance for the model’s concept.
• There was no readily available training data to tag roof types for forecasting damage intensity.
• Many disaster-prone areas are off the technology grid with low mobile penetration and network connectivity.

Sources

• gramener.com — vendor self-report, 2023-07-06