dunnhumby Science blog

Looking at how graph models can deal with the added complexity of time.

A non-technical introduction to the world of graph data structures and the graph neural networks built to analyse them.

Designing a pipeline strategy that simplifies deployments in a multi-component domain

Making deployments a matter of configuration, not code implementation.

Can Newton’s idea of gravity be used to estimate grocery store revenue?

How we halved the runtime of a data-hungry application in Spark

The supermarket can be a daunting place. How do we know what to buy? Here I summarise some key findings from my Psychology PhD @ UCL.

How we track thousands of data science projects, by Alapon Sen

Forecasting customer demand with probabilstic models

David Hoyle, Price & Promotion Science, dunnhumby

How we’re supercharging Customer Data Science using programmable chips

Software does not build itself, people do — here’s how they will impact the success of your business technology transformation.

How close are we to having robot colleagues and managers? How about the concept of a wise machine?

By George Jones, Software Engineer at dunnhumby

Discover the steps taken to evolve our Media’s design system to a managed one offering our teams a clear modular way to build applications