NWCA Blog

How We Built a Property Comps API Covering 11 Global Markets

Property comparable sales data is fragmented. Every country has its own government registry, its own data format, its own quirks. If you are a proptech company, real estate investor, or AI agent that needs comps across multiple markets, you are looking at 11 different integrations.

We unified all of them into one REST API.

The Markets

MarketSourceTransactionsApp
United KingdomHM Land Registry31M+property.nwc-advisory.com
FranceDVF / DGFiP8.3M+property-fr.nwc-advisory.com
New York CityNYC Dept of Finance51K+property-nyc.nwc-advisory.com
Singaporedata.gov.sg (HDB)973K+property-sg.nwc-advisory.com
DubaiDubai Land Dept1.3M+property-dxb.nwc-advisory.com
ChicagoCook County180K+property-chi.nwc-advisory.com
MiamiMiami-Dade County81K+property-miami.nwc-advisory.com
PhiladelphiaCity of Philadelphia48K+property-phl.nwc-advisory.com
ConnecticutCT Open Data155K+property-ct.nwc-advisory.com
IrelandProperty Price Register27K+property-ie.nwc-advisory.com
TaiwanMinistry of Interior23K+property-tw.nwc-advisory.com

Architecture

Each market runs its own FastAPI backend with a local SQLite database. A unified gateway at api.nwc-advisory.com routes requests by the market parameter to the right backend:

GET /v1/comps?market=nyc&zip_code=10001&limit=10&months=12
GET /v1/stats?market=uk&postcode=SW1A1AA
GET /v1/trends?market=fr&code_postal=75001

The gateway normalizes responses into a consistent schema regardless of which market you query. Every response includes the same fields: price, date, address, property type, area, and price per unit.

Why Government Data

Every transaction in our database comes from an official government registry. This matters because:

The Hard Parts

Location parameters differ by market. UK uses postcodes (SW1A1AA), France uses code postal (75001), NYC uses ZIP codes (10001), Dubai uses area names (Dubai Marina). The gateway auto-detects which parameter to route based on the market.

Database performance at scale. The UK database has 31 million rows in an 8.7GB SQLite file. A radius search around a dense postcode like Norwich NR1 used to take 168 seconds. Compound indexes on (postcode, property_type, date DESC) brought it down to under 5 seconds.

Currency and unit normalization. UK reports price per square foot, France uses price per square meter, Singapore uses price per square foot but in SGD. The API returns raw local values plus a normalized price_per_unit field.

AI Agent Integration

We also built an MCP server so AI agents (Claude Desktop, Claude Code) can query property data directly. The server exposes three tools:

Explore the API - Swagger Docs

Free tier: 50 requests/month, 10 results per search. No credit card required.

About NWCA

New Way Capital Advisory builds financial analytics tools for wealth managers and investors. Beyond property comps, our platform includes: