Data tools that run every day.
We use every service on this page ourselves: Cercand's price observatory and APIs are the proof.
Web and API development
Custom web applications and APIs: from an internal tool for looking up data to a public API with access keys.
- Source code and documentation
- Deployment on your server or ours
- Maintenance and evolution
This website, in four languages, and the two public price APIs for supermarkets and pharmacies.
Data collection
Automatic collection of data published on the web, integrated with the sources you already have: spreadsheets, databases or third-party APIs.
- Clean data in a database, a CSV or an API
- Scheduled collection at whatever frequency you need
- Alerts when a source changes
Eight supermarket and pharmacy websites, collected every night.
Automation
Repetitive tasks turned into processes that run on their own, with a log of every run and an alert when something fails.
- Scheduled, documented processes
- A log of every run
- Alerts when something fails
The observatory's nightly refresh: it reads millions of prices, recomputes the index and publishes the figures with nobody touching it.
Data analysis
Answers to specific questions about your data or the market's, with the right method and its limits spelled out.
- Results with the methodology explained
- Reproducible data and calculations
- Conclusions in plain language
A chained Jevons price index, the method statistics offices use, across four supermarket chains.
Reports and dashboards
The data you decide with, laid out to be read in a minute and to keep itself up to date.
- A weekly or monthly report
- A private web dashboard
- Alerts when a figure moves
The price thermometer and the pharmacy comparison on the home page, updated every night.
Data consulting
Before building anything: what data you really need, where it can come from, what it costs to maintain and what it can tell you.
- Source and feasibility assessment
- Proposed architecture and method
- A phased plan
Telling a price change from something that only looks like one: products a shop relists under a new code, or prices that switch from per unit to per kilo.
How we work
We understand the question
What you want to know or automate, and what you will use it for.
A prototype on real data
A small first version, on real data, to check that it answers the question.
Going live
Deployed, documented and running on its own.
Maintenance and evolution
Websites change. When they do, we take care of it.
Is there data you can't get, or data you can't make sense of?
Tell us what you need and we'll tell you how we would do it.