The task
The agency sells overseas real estate and wanted to show clients properties from several foreign platforms in one catalog and in one language. The problem: different countries, languages, currencies, area units and listing formats — consolidating that by hand is impractical.
What we did and what data we collected
We set up listings collection from several foreign platforms: price, area, property type, location, description, photos. We normalized the data (currencies, area units, property types) and translated the listings into the agency's language. The output — a single catalog fit both for showing to clients and for internal analytics.
Challenges and how we solved them
First, multilingual sources and diverging formats. Every platform has its own language and structure, so we first mapped listings to a common field set and only then translated — that kept the translation consistent.
Second, translation quality in niche vocabulary (real estate terms). We tuned the translation for the domain and kept the original alongside so a manager could double-check.
Third, currencies and units. Prices and areas were converted to comparable values while preserving the original figures. Duplicates across platforms were removed so the catalog would not swell with copies.
The result
The agency received a single catalog of overseas properties in one language, with comparable prices and specifications — no more manual translation and country-by-country spreadsheet stitching.
Services in this case study: Real estate scraping · Automated product listing translation · AI Data Processing · Data for the Real Estate Market
Useful reading: Scraping dynamic sites · Data normalization