Local data
Customer review scraping
An average rating tells you nothing: what informs is its slope, and what the texts behind it say. Collecting reviews means retrieving the full history rather than today's figure.
Data we extract
- Rating and date of each review
- Public verbatim and source language
- Owner response
- Average rating and its curve over time
- Distribution of ratings by star
- Recurring themes after analysis
- Review volume and posting rhythm
Example use cases
- Reputation monitoring across a retail network
- Sentiment analysis and recurring themes
- Comparison with nearby competitors
- Detecting a quality incident
Framework & ethics
We work with publicly accessible data, at a controlled pace so as not to disturb the sites we consult. Defining the uses is your responsibility, and we advise you on it.
Other kinds of site we collect
Businesses and points of interest
Local data
A business record looks simple until you want ten thousand of them, current, free of duplicates and geocoded.
Property listings
Real estate
Listing portals publish tens of thousands of properties every day, and pull them just as fast.
Estate agency networks
Real estate
A network's own site does not publish what a portal publishes: it shows a portfolio, a fee scale, a web of branches and the negotiators who run them.
Vehicle listings
Motoring
A used car's price means nothing on its own: it hangs on the model year, the mileage, the trim and who is selling.
Valuations and spec sheets
Motoring
A valuation and a listing say different things: one gives a reference value, the other an asking price.
Job postings
Hiring
The same role appears on several sites under three different titles, then vanishes without notice.
Salaries and employer reviews
Hiring
What a role pays, and what people think of the employer, are not in the job ad.
Marketplaces
E-commerce
On a marketplace a product has no single price but as many prices as it has sellers, and the one that wins the basket changes during the day.
All kinds in the same sector: Local data scraping
Need data of this kind?
Tell us about your project: sources, fields, volume and frequency. We'll come back with a clear proposal.