How our models work
Gresp is powered by MPNN_LSTM, a temporal graph neural network that combines daily or monthly regional values with a region-connectivity graph to forecast near-term demand or risk per region — across epidemic surveillance, pharmacy demand, mobility demand, and more.
01
Ingest
Collect uploaded data, surveillance feeds, or operational indicators.
02
Transform
Clean, align, and engineer temporal features for forecasting.
03
Predict
Run trained models to estimate near-term demand, risk, or intensity.
04
Act
Present actionable outputs to operational teams and decision-makers.
Bring your own data
Upload
Have your own regional data and a connectivity graph? Upload it and Gresp will train and forecast on it directly.