London, UK
FoodBridge is a research prototype: a coordination layer for food redistribution in London that forecasts surplus before it happens, allocates it using fairness criteria and consolidates pickups into efficient routes. You just say what you have left over.
Results from a calibrated simulation over London's real geography (MSc study, Northumbria University, 2026).
Every day, London shops throw away perfectly good food while food banks and community centres need it just a few kilometres away. The problem isn't willingness — it's coordination. FoodBridge solves it with three pieces of technology.
Machine-learning models (XGBoost and a neural network) predict how much surplus each shop will have, and in which time slot — up to 25% more accurately than estimates based on historical averages. Pickups get planned before the food is even left over.
An optimiser decides which donation goes to which organisation, weighting each neighbourhood's need (the official Index of Multiple Deprivation), not just proximity. The result: more food reaches where it's needed most, without driving a single extra kilometre.
The day's pickups are grouped into optimised van routes that cut close to 70% of the kilometres compared with direct shop-by-shop trips. Less fuel, less CO2, more deliveries per driver.
No phone calls, no WhatsApp coordination, no vans kept waiting. Three steps.
30 seconds: what you have left over, how many kilos and when it's ready. The system already knows where you are and when you close.
The optimiser proposes the recipient and you choose: a driver already en route stops by, a bike volunteer comes… or you walk it over if it's close. And if anything fails, the system reassigns automatically — there is always someone responsible.
Follow the delivery live and build your track record: kilos donated, meal equivalents and CO2e avoided. Your surplus stops being waste and becomes data… and dinners.
FoodBridge doesn't promise — it measures. In an experiment over London's real geography (shops, organisations and official deprivation by neighbourhood) with simulated surplus calibrated to sector figures, it was benchmarked against today's operating practices — fixed shop lists, zone depots and the classic first-come-first-served.
Scaled city-wide, the system would recover around 7,800 additional tonnes a year: ≈ 18.6 million meals, £30.7M in recovered food value and 32,700 t of CO2e avoided.
Results from a simulation calibrated with open data and statistically validated (MSc study, Northumbria University, 2026).
FoodBridge London grew out of MSc Big Data and Data Science Technology research (Northumbria University) by David Borrero Diaz: a decision support system that combines demand forecasting, geospatial analytics and route optimisation to reduce retail food waste. Built with the Design Science Research methodology and calibrated with open data: OpenStreetMap, the English Indices of Deprivation 2019 and the London Datastore.
This site is a working demo: it simulates a full operating day (07:00–20:00,
accelerated) using the real assignments computed by the optimiser in the experiment.
Tip: open coordinator and shop in two tabs — your donation appears live in the control room.
Sign in as a shop and donate your first surplus, or open the control room and watch the day unfold in real time.
No donations awaiting assignment.
The day starts at 07:00…