Heat Pump Ready?¶

Client: Matt Millar, HomeServe EMEA.

Millions of UK homes will need to swap gas boilers for heat pumps, but many owners have no idea whether their home is suitable, and a professional survey costs time and money. The key questions are how much heat each room loses, and whether the existing radiators can deliver enough warmth at the lower water temperatures that heat pumps use.

Your task is to build a phone app that lets a householder walk through their home, scanning rooms and photographing radiators, windows and the loft, and then tells them whether their home is ready, nearly ready or not yet ready for a heat pump, and what would change that.

The team must infer insulation from indirect clues and public data, combine this with the home's actual gas use, and be honest about uncertainty.

HomeServe can provide professional scanning tools and real evaluations to compare against the app's outputs.

This project is a strong candidate for machine learning, with scope for several different techniques, for example recognising radiators, windows and insulation from photos and scans, and estimating heat loss with honest uncertainty.