# Centralising property listings from multiple sources

> A real estate agency in Turkey wanted to display an up-to-date property catalogue without manually re-entering every listing on its site.

- Source canonique : [https://allaux.fr/en/realisations/collecte-annonces-immobilieres](https://allaux.fr/en/realisations/collecte-annonces-immobilieres)
- Langue : EN
- Dernière mise à jour : 2026-09-30

## The context

The agency's listings were spread across several places: some entered directly by agents, others picked up from partners. The agency's showcase site, meanwhile, was updated by hand, with a frequent lag between the real state of the market and what was actually visible online: a sold property sometimes stayed listed for several days, while a newly available one took time to appear.

The goal was to build a collection mechanism that gathers the existing listings, normalises them into a shared format, and publishes them on the showcase site without manual re-entry, with minimal lag.

## What I did

1. **Identified the sources** — Mapped out the different places where listings actually existed, along with their respective formats and update frequencies.
2. **Normalised the data** — Built a shared property model — floor area, type, location, photos, price — that each source is converted into.
3. **Detected removed listings** — Built a mechanism that spots when a property no longer appears in a source and removes it from the site automatically, rather than leaving it listed by default.
4. **Automated publishing** — Updated the showcase site regularly and automatically from the centralised catalogue, with no manual intervention for each change.

## Outcome

> The showcase site reflected the real state of the property catalogue with a much shorter lag, without anyone having to re-enter a listing that already existed elsewhere.

## FAQ

### How are property photos handled?

They're retrieved and matched automatically to the corresponding listing during normalisation, with a minimal quality check before publishing.

### What happens if a source becomes temporarily unavailable?

The centralised catalogue keeps the last known data rather than emptying the site, until the source becomes reachable again.

### Can this mechanism be adapted to new sources later on?

Yes, adding another source just means writing its conversion to the shared model already in place, without reworking the whole system.

### Can the update frequency be adjusted?

Yes, it's set according to how fast the relevant market actually changes, with no need for further development to change it.