Skip to main content

Data science in government

Posted by: , Posted on: - Categories: Open Innovation Team, Skills

What is data science?

Data science sits at the intersection of statistics, computer science and deep business or subject knowledge.

A Venn diagram with three overlapping circles. Occupying the largest circles are: "coding skills", "maths and statistics" and "substantive expertise". Where "substantive expertise" and "coding skills" overlap, there is "danger". Where "coding skills" and "maths and statistics" overlap, there is "machine learning". Where "maths and statistics" and "substantive expertise" overlap, there is "traditional research". The danger section represents the fact that data science techniques can be used incorrectly if they are not supported by a rigorous understanding of mathematics

This data science venn diagram, adapted from  Drew Conwayvisualises how the disciplines that make up data science interact.

The intersection between:

  • Maths and substantive expertise is traditional research - the systematic exploration of a subject
  • Coding and maths is machine learning - the automation of statistical processes for ‘learning’
  • Coding and substantive expertise is a danger zone - data science techniques can be used incorrectly if they are not supported by a rigorous understanding of mathematics

A data scientist may not be an expert in all of these areas. However, they will require a working knowledge of all disciplines and will often work closely with other experts in a multidisciplinary team.

What can you do with data science?


Data visualisation is presenting data in a digestible and interpretable way. It helps to understand what the data shows and to communicate insights.


A mind map showing the kinds of things data science is concerned with. Outwards pointing arrows point to the words, "classification", "anomaly detection", "regression", "reinforcement learning" and "clustering"

As summarised by Microsoft Azure, there are five basic types of question that data science can be used to answer:

Question Type of Algorithm Description
Is it A or B? Classification Grouping data into predefined categories
Is this unusual? Anomaly detection Identifying unexpected or unusual events or behaviours
How much or how many? Regression Making numerical predictions
How is this data organised? Clustering Understanding the structure of the data
What to do next? Reinforcement learning Making decisions within an environment where correct actions are rewarded and future decisions improve

How is data science being used in government?

The adoption of data science is enabling government to unlock the value of the data it holds.

It is being used to:

The Open Innovation Team is identifying opportunities for increasing academic engagement to support data science projects across government. If you would like to get in touch with us about our data science work you can email us or follow us on twitter @openinnovteam.

If you want to explore data science further, this post from the Government Digital Service outlines their approach to data science projects and this blog gives some suggestions on how to get started.

Sharing and comments

Share this page