Inspiration
With the 2030 deadline fast approaching, UK landlords are under immense pressure to hit EPC rating 'C'. However, the path to compliance is a "planning minefield." Landlords often spend thousands on professional fees for upgrades—like external wall insulation—only to have them rejected because the property sits in a Conservation Area. We wanted to build a tool that gives landlords the "Safe Path" to compliance using data, not guesswork.
What it does
Retrofit ROI is an AI-powered decision engine. A user enters a postcode, and the system instantly orchestrates three data streams:
EPC API: Fetches the current rating and assessor-recommended upgrades.
Spatial Data: Checks the property against Geointersects boundaries for Conservation Areas. It then uses Gemini 1.5 Pro to synthesize this into a "Safe vs. Danger" action plan.
How we built it
We built a full-stack application using a high-performance architecture:
Backend: Node.js and Express (orchestrating parallel API calls).
AI: Google Gemini 1.5 Pro for intelligent data synthesis and planning advice.
Database: MongoDB Atlas for geospatial indexing and result caching.
Frontend: A modern React dashboard built with Vite and Tailwind CSS, focused on high-contrast "Traffic Light" status reporting.
Spatial Logic: Implementation of coordinate-based containment checks for UK conservation zones.
Challenges we ran into
The biggest technical hurdle was Data Orchestration. Fetching data from three different sources (EPC, and Spatial files) and feeding it to an LLM in a single request cycle can be slow. We solved this by implementing async parallel processing and a MongoDB caching layer, which reduced subsequent search times from 5 seconds to sub-500ms.
Accomplishments that we're proud of
We are incredibly proud of our Planning Success Predictor. By leveraging data, we aren't just telling landlords what to do, but the probability of it being approved. We successfully integrated complex Geointersects spatial data with real-time AI logic to turn a complex legal problem into a simple "Green/Red" UI.
What we learned
We learned the power of "Context-Augmented LLMs." Gemini is powerful on its own, but when fed specific, hyper-local data from EPC records, it provides professional-grade planning advice that is specific to a single street, rather than generic AI advice.
What's next for Retrofit ROI
The next step is to integrate a "ROI Calculator" that factors in the cost of the "Safe Path" upgrades versus the projected increase in property value. We also aim to add an automated "Planning Application Generator" that uses Gemini to pre-fill council forms based on the "Safe" upgrades identified by our system.
Log in or sign up for Devpost to join the conversation.