Your partner for complex spatial data challenges.
A close-knit team in Utrecht that likes hard problems. You won't catch us chasing the latest AI trends; empty buzzwords make us twitchy. We'd rather build algorithms that work and help you make better decisions.
IRIS is a location strategy platform, built for retail and franchise networks.
IRIS helps you make better location decisions.
A failed opening is wasted money. IRIS helps you learn from the past so you can make better decisions in the future. We don't just give you a forecast, we back it up. We are transparent about the uncertainty margins and the variables driving the prediction.
1 Validated out-of-sample against 2020-2025 opening cohorts.
Visit irisplatform.nl →Whether we're anonymising millions of documents for municipalities or designing algorithms to understand retail areas, we don't stop until the infrastructure is rock solid.
Evaluating and comparing international expansion sites without relying on local, fragmented, and disconnected data models.
Reducing reliance on manual, error-prone field data collection for large-scale geographic property assessments.
Automating the ingestion, cleaning, and structuring of millions of raw geographical coordinates into production-ready predictive models.
not quarters Direct to production
- Retail risk modelling A single risk score for every retail property in the Benelux.
- Real-time footfall data Smartphone pings combined with IR camera data for accurate monitoring.
- Company lookalikes Identifying similarity between companies based on websites and social media.
Both. We sell IRIS as a subscription. If your challenge is too specific for an off-the-shelf product, that's custom work, and you work directly with the people writing the code, not a layer of consultants in between.
No. No months-long discovery phases that end in a slide deck or report. We'd rather get started right away and build a proof of concept.
Depends on the project, of course. Usually we deliver a first prototype within a few weeks.
More often than you might think. We're honest about our limits, like model uncertainty or history that's too thin. And we believe the combination of algorithm and on-the-ground knowledge leads to the best results.
We validate out-of-sample and track accuracy after the fact. We've been wrong before; when we are, we document it and fix the pipeline. That's how the algorithms keep getting better.
Partly, which is why most of our effort goes into which relationships actually hold up. A correlation that collapses the moment the market turns is no basis for a forecast.
Ready to start? Drop us a line.
info@thebigdatacompany.nl →The Big Data Company B.V.
Princetonlaan 6 · 3584 CB Utrecht, Netherlands
+31 30 899 9477