Frequently Asked Questions
The platform combines regional data, temporal features, and machine learning signals to estimate short-term demand or risk and support earlier operational decisions — the same underlying model powers our epidemic forecasting demo and our pharmacy and mobility demand demos.
Gresp (Graph Response) is a Graph Neural Network framework, originally built for predicting disease trends and since extended to regional demand forecasting across other domains.
Upload a zip with a labels.csv (a region-level value per day — case counts, prescribing volume, demand, etc.) and a graphs/ folder (one region-connectivity CSV per day). The Upload page checks your files against the format the model expects and tells you exactly what to fix.
No. The England COVID-19 demo was the original use case, but the same temporal graph neural network now also powers live demos for regional pharmacy prescribing demand and NYC taxi zone demand, and has been tested on retail replenishment and energy-grid demand as well. Try the other datasets on the Demo page.
Yes — sign up with email/password or Google, and your session is remembered across visits.
Your uploaded data is processed automatically to train and run your forecast — it is never manually reviewed by anyone at Gresp. It is stored on Google Cloud Storage, encrypted at rest, and the raw files you uploaded are automatically deleted once training completes; only your results and trained model are kept. See our Privacy Policy for the full details.