PropertyOS · Index 01

Europe's propertymarkets, quantified.

ML-powered rental pricing, 34 economic datasets and A–F quality grades across all 27 EU countries — built on official ECB and Eurostat data, with model performance published per country.

—
EU countries
—
Indicator observations
0
Economic datasets
—
Avg model R²
Model fit (R²)Loading country models…Latest model per country
02 · The problem

Building this in-house takes eighteen months. Scraping it yourself is a liability. We run the pipeline — you ship the product.

03 · The pipeline

From official sources

to one API call.

Three stages, fully automated. The figures below are counts from the live catalog.

01

Official market data

Official series from the ECB and Eurostat, with OECD and World Bank context datasets, loaded into one normalized schema. Every dataset carries its source and a data-through date.

ECBEurostatprovenance per dataset
34Datasets
02

Measured models

One Ridge model per country, retrained roughly every two weeks. Every training run is logged and cross-validated R², MAE and RMSE are published per country.

per-country modelslogged retrainingpublished metrics
27Country models
03

Production-ready API

JSON endpoints with validated request and response contracts, rate limits and health checks. Current service health is published as queryable endpoints.

health checksrate limitstyped contracts
LivePublished status
04 · Under the hood

Watch the data move

1Ingest

Official sources, one schema

ECB rates and Eurostat indices land in one normalized schema, with the source and observation date kept on every value.

2Validate

Quality is graded, not assumed

Every dataset is scored A–F for completeness, recency and provenance. Values that are impossible for their unit are withheld and flagged instead of shown.

3Model

Twenty-seven country models

Ridge models per country with sample-based regularization, retrained on a regular cycle. Cross-validated R² is published for every country, including the weaker ones.

4Serve

Served in one API call

One JSON request returns a 0–100 FairRent Score, an A–F fairness grade, the market price per m² and the monthly difference against market rent.

05 · The API

One call.
Every market.

Pricing intelligence, 34 economic datasets and transparent model performance behind plain JSON endpoints.

JSON over HTTPSPlain POST/GET endpoints, no SDK required
Validated contractsRequests and responses checked against published schemas
Rate limitsPer-client limits with Retry-After headers
Health endpointsCurrent database and model status you can query
curl -X POST /api/fairrent/calculate-score \
  -H "Content-Type: application/json" \
  -d '{
    "country": "DE",
    "rent": 1500,
    "size": 60,
    "quality": 5,
    "rental_type": "entire_property",
    "furnishing_status": "unfurnished",
    "utilities_included": false
  }'
06 · Coverage

Twenty-seven markets,
measured in public.

The latest model for every EU country, sorted by cross-validated R² — the weaker fits are listed too.

CodeCountrySamplesModel fit (R²)Status

Loading country models…

—
Countries covered
—
Trained models
—
Indicator observations
—
Average R²
08 · Start

Start building,
today.

Book a Demo
Free accountNo credit cardEU27 coverage