HomeMatch 🏠
Find a home based on where your life happens – not just where you think you should live.
HomeMatch is a commute-based housing search for Dublin. Instead of asking "Where do you want to live?", it asks "Where do you need to go?" – then ranks homes by how well they fit your budget, your commute, your public transport options and your lifestyle, and explains why each one was recommended.
Built for the Hack for Humanity – Local Track hackathon. 📑 View the pitch deck (PDF)
The problem
Property sites make you pick an area first, then filter by price and bedrooms. But most people don't need one specific neighbourhood – they need to reach work, school, family and childcare easily and affordably, with or without a car.
Because people instinctively search as central as possible, demand concentrates in a few areas and prices rise disproportionately, while well-connected homes further out go unseen.
The result: a suitable home often exists somewhere you never thought to look, because it appears "too far" on a map even though it sits next to a tram stop, or a home that looks close has no useful transport at all.
The solution
| Traditional search | HomeMatch |
|---|
| Starting point | Area → Properties | Your destinations → Properties |
| Distance | Straight-line / map distance | Real door-to-door commute time by your chosen modes |
| Results | Strict filters (hide anything that misses) | Match Score that surfaces good trade-offs |
| Transparency | A list | "Why we recommended this" – strengths and trade-offs |
Example: a €1,800/month apartment that isn't walkable to your office but is beside a Luas stop and needs no car can outrank a closer one with poor connections.
Who it's for
Renters and buyers new to a city, students, young professionals, families, people changing jobs, people without cars, and hybrid workers – including households with several people going to several places.
Features
- Multi-destination preferences – add workplaces, university, school, family, etc., each with a type, visits per week and importance (low / medium / high).
- Transport modes – car, train, bus, bike, walk.
- Match Score (0–100) per property, with sub-scores for affordability, commute, transport, property fit and lifestyle.
- Transparent, no-AI scoring – a simple weighted algorithm you can inspect; no wasteful use of AI.
- Smart categories – Best overall, Best value, Best commute, Worth considering.
- Plain-language explanations – strengths, trade-offs and how far each result is over/under budget or commute limit.
- Interactive map (Leaflet + OpenStreetMap) next to ranked property cards.
- Property detail page with per-destination commutes and nearby kindergartens, schools and grocery stores (within 1 km).
- Real data – Dublin rental listings from Daft.ie, NTA GTFS stops/lines, OpenStreetMap routing.
- Polished "Night Line" dark UI – scroll-scrubbed Luas-door hero, pixel-art DART loading screen and a few hidden easter eggs (see
docs/whimsy.md).
How the Match Score works
Every property is scored rather than hard-filtered:
| Component | Weight |
|---|
| Affordability | 35% |
| Commute (weighted by destination importance × visits/week) | 30% |
| Transport connectivity (from GTFS stop/line data) | 15% |
| Property requirements (type, bedrooms, …) | 10% |
| Lifestyle (nearby amenities) | 10% |
Commute estimates
For each destination, HomeMatch builds candidate journeys for the modes you picked and times each leg on OpenStreetMap:
- Walk / cycle: OSM foot/bike route.
- Drive: OSM car route × 1.4 (OSRM assumes empty roads) + 5 min parking.
- Bus / Luas: walk to a stop + half the peak headway + OSM road time × 1.6 (bus) / 1.3 (Luas) + walk from the stop.
- DART / rail and journeys needing a change are shown as estimates.
The first search for a new set of destinations takes ~10 s (requests are spaced out for the shared OSRM server); repeats are served from cache.
App overview
| Route | Purpose |
|---|
/ | Landing page with the Luas-door hero |
/preferences | Budget, bedrooms, max commute, transport modes, destinations |
/results | Ranked property cards + map |
/properties/[id] | Property detail, "why it matches", commutes, amenities |
POST /api/match | Takes preferences, returns ranked MatchResult[] (validated, JSON-only, 16 KB cap) |
GET /api/amenities/[propertyId] | Precomputed nearby amenities |
Tech stack
- Framework: Next.js 15 (App Router), React 19, TypeScript
- Map & UI: Leaflet / react-leaflet, framer-motion, lucide-react, plain CSS design tokens
- Backend: Next.js route handlers – no separate server or database
- Security: strict CSP and security headers in
next.config.mjs, input validation in lib/validate.ts
Data sources
| Data | Source | When |
|---|
| Listings (11 Dublin rentals + 2 sample purchase listings) | Daft.ie – coordinates, rooms, facilities in lib/data.ts | static |
| Stops and which lines serve them | NTA GTFS timetable → data/generated/transit.json | build time |
| Journey times | OSRM table service, batched + cached | request time |
| Kindergartens, schools, groceries | OpenStreetMap Overpass → data/generated/amenities.json | build time |
| Destination coordinates | OpenStreetMap Nominatim (1 req/s, cached) | request time |
| Map tiles | OpenStreetMap via Leaflet | browser |
Getting started
Requires Node ≥ 22.18 (the data scripts import lib/data.ts using Node's built-in TypeScript support). No API keys needed.
npm install
npm run dev # http://localhost:3000
The generated data in data/generated/ is committed, so the app runs straight away.
| Script | What it does |
|---|
npm run dev | Start the dev server |
npm run build / npm start | Production build / serve |
npm run lint | Lint the project |
npm run prepare-data | Refresh GTFS transit data (~1 min) and OSM amenities (a few min) |
npm run build:transit / npm run build:amenities | Run either data step on its own |
Project structure
app/ Pages and API routes (landing, preferences, results, properties, /api/*)
components/ Map, header/footer, commutes panel, UI + whimsy components
lib/ Matching engine (matching.ts), listings (data.ts), validation,
transit/ (OSRM routing, GTFS connectivity, geocoding)
data/generated/ Prebuilt transit and amenity data
scripts/ build-transit.mjs, fetch-amenities.mjs
docs/ design-system.md ("Night Line"), whimsy.md
public/hero/ Hero imagery
Roadmap
- Real-time listings through integration with property platforms like Daft.ie
- Environmental impact – reward the benefit of not needing a car when evaluating a property
- Additional factors: local safety, air quality and environmental information (with sparing, sustainable use of AI)
- Hybrid-working schedules and per-user weight customisation
- Healthcare and accessibility requirements
- Commute cost, car ownership and parking cost comparison
- Saved searches, alerts and side-by-side comparison
Impact
Better search won't solve housing supply, but it can improve housing discovery. HomeMatch does the transport analysis for you, so people can find well-connected homes in areas they hadn't considered – without researching every district of the city.
HomeMatch is Skyscanner for housing. Don't search neighbourhoods. Search for the home that fits your life.
Team
Built by Alper Ergüne, Guillermo Garcia and Kacper Kotwica for the Hack for Humanity Local Track.
License
No license has been specified yet.
