Data Journalism for Beginners: A UK Reporter's Guide

This guide is for reporters who have little or no experience working with data but want to add data skills to their toolkit. It does not require coding ability. If you can use a spreadsheet, you can do data journalism.

Data journalism is one of the most in-demand skills in UK newsrooms, yet many reporters feel intimidated by it. The truth is that the most impactful data stories often rely on nothing more complicated than a spreadsheet and a good question. The Bureau of Investigative Journalism, the BBC Shared Data Unit, and regional newspapers across Britain are all producing excellent data journalism — and with the right approach, so can you.

Where to Find UK Data

The UK has some of the best publicly available datasets in the world. Knowing where to look is half the battle. Here are the key sources every UK reporter should bookmark:

Office for National Statistics (ONS): The UK's largest independent producer of official statistics. Covers population, employment, health, crime, housing and much more. The ONS website allows you to download data in CSV and Excel formats, and its NOMIS service provides detailed labour market data by local authority.
data.gov.uk: The UK government's open data portal, hosting over 50,000 datasets from across government. Includes everything from school performance data to environmental monitoring.
NHS Digital (now NHS England): Health data for England including hospital performance, prescribing data, and workforce statistics. Prescribing data is particularly rich — you can see exactly which drugs are being prescribed in every GP practice in England.
Police.uk: Street-level crime data for England, Wales and Northern Ireland. Can be downloaded as CSV files for analysis.
Companies House: Free access to company filings, accounts and director information. Essential for investigative reporting.
Local authority data: Council spending, planning applications, FOI disclosure logs and more. Quality and accessibility vary enormously between councils.

Devolved Statistics: Scotland, Wales, and Northern Ireland

Many reporters default to ONS data and miss the devolved statistical agencies, which hold the most granular data for stories in Scotland, Wales, and Northern Ireland. Each agency has its own portal and methodology, so always read the accompanying notes before making cross-country comparisons.

National Records of Scotland (NRS): Population, births, deaths, marriages, and census data for Scotland. NRS runs Scotland's Census separately from ONS and publishes its own midyear population estimates.
Northern Ireland Statistics and Research Agency (NISRA): The official statistical agency for Northern Ireland. Covers population, economy, health, and justice. NISRA conducts its own census in conjunction with NRS on a different cycle from England and Wales.
StatsWales: The Welsh Government's statistics portal. Covers education, health, economy, transport, and environment data for Wales, often at a more granular local authority level than ONS outputs.
Scottish Government Statistics: Supplementary data on Scottish policy areas not covered by NRS, including justice, housing, and environment. Published at statistics.gov.scot.

Essential Spreadsheet Skills

You do not need to learn Python or R to do effective data journalism. Google Sheets or Microsoft Excel will handle most reporters' needs. The key skills to master are:

Sorting and filtering: The most basic but powerful technique. Sort data to find the highest and lowest values; filter to focus on specific categories or time periods.
Pivot tables: These allow you to summarise and cross-tabulate large datasets. For example, you could use a pivot table to show total crime by type for each local authority.
VLOOKUP and INDEX/MATCH: These functions let you combine data from different sources. For example, matching crime data with population data to calculate crime rates per capita.
Percentage change: Essential for putting numbers in context. The formula is (new value - old value) / old value * 100.
Basic statistics: Mean, median, mode and standard deviation help you understand the shape of your data and identify outliers.

Google offers free training through its Google News Initiative data journalism courses, and the European Journalism Centre provides excellent online resources specifically for journalists.

Visualising Data with Datawrapper

Datawrapper is a free tool used by newsrooms including the BBC, the Guardian and the Financial Times. It allows you to create publication-ready charts, maps and tables without any coding knowledge. To get started:

Create a free account at datawrapper.de
Paste or upload your data — Datawrapper accepts CSV, Excel and Google Sheets links
Choose a chart type — bar, line, pie, scatter, map or table
Customise the design, add annotations and source credits
Publish and embed the chart in your article

Other useful free visualisation tools include Flourish (excellent for animated and interactive visualisations), RAWGraphs (for unusual chart types) and Google Data Studio (for dashboards). The key principle is to choose the simplest chart type that tells your story clearly — a well-designed bar chart is almost always better than a flashy but confusing 3D visualisation.

Using FOI to Get Data

Not all useful data is publicly available. The Freedom of Information Act 2000gives you the right to request data held by public authorities. FOI is one of the most powerful tools in the data journalist's toolkit, but it requires patience and strategy.

Ask for data in a structured, machine-readable format (CSV or Excel) rather than PDF.
Be specific about the fields, time period and geographic scope you need.
Check the authority's disclosure log first — the data may have already been released.
If you need the same data from multiple authorities, use WhatDoTheyKnow.com to send batch requests efficiently.
Be prepared for refusals citing cost — if your request would exceed the cost limit (18 hours of work for central government, 24 hours for local authorities), narrow your scope.

Storytelling with Numbers

Finding interesting data is only half the job. The other half is turning it into a story that readers care about. The best data journalism combines rigorous analysis with compelling narrative:

Lead with the human impact. A statistic about hospital waiting times becomes a story when you find the patient who waited 18 months for surgery.
Provide context. Raw numbers are meaningless without comparison. Is this number going up or down? How does it compare to other areas? What is the national average?
Localise. National datasets can be broken down to local authority, parliamentary constituency or postcode level. Local angles drive engagement.
Be honest about limitations. Every dataset has gaps, inconsistencies and caveats. Acknowledge them in your reporting.
Show your working. Where possible, make your data and methodology available to readers. Transparency builds trust.

Common Pitfalls to Avoid

Correlation is not causation. Just because two trends move together does not mean one causes the other.
Small sample sizes. Be cautious about drawing conclusions from small numbers. A 100 per cent increase sounds dramatic, but not if it means going from one incident to two.
Cherry-picking. Do not select only the data points that support your hypothesis. Good data journalism follows the evidence wherever it leads.
Misunderstanding rates and ratios. Always check whether you are looking at raw numbers or rates. Crime may be rising in absolute terms but falling per capita if the population has grown.

Practical Checklist

Before publishing a data-driven story, run through these checks:

Have I identified the correct statistical agency for the nation(s) covered by my story (ONS, NRS, NISRA, or StatsWales)?
Have I read the dataset's methodological notes and checked for any known caveats or data quality warnings?
Have I checked whether I am comparing like with like across years or geographic areas (same definitions, same population base)?
Have I converted raw counts to rates or percentages where population size differs between areas?
Have I had a statistician, academic, or data-literate colleague review my analysis before publication?
Does my chart or visualisation accurately represent the data without misleading scales or truncated axes?
Have I credited the data source clearly in both the article and the chart?
Have I found at least one human story or expert quote to give the numbers context and reader relevance?

Common Mistakes

Using ONS data for Scotland-specific stories. ONS covers England and Wales. For Scotland, use NRS; for Northern Ireland, use NISRA. Cross-country comparisons require careful harmonisation.
Requesting data as a PDF. Always specify CSV or Excel in FOI requests. PDF data requires manual re-entry or unreliable optical character recognition.
Publishing a chart without a source credit. Every chart must credit its data source clearly and, where possible, link to the underlying dataset.
Treating provisional statistics as final. Many datasets — particularly crime, health, and economic data — are initially published as provisional estimates subject to revision. Label them as such.
Ignoring the disclosure log before filing an FOI request. Many frequently-requested datasets are already published. Checking the disclosure log first saves weeks of waiting.
Choosing a chart type for visual impact rather than clarity. A pie chart with ten segments and a 3D effect communicates nothing. A simple bar chart ranks data clearly.

Red Flags

A press release that cites a striking percentage change without showing the underlying numbers — check the sample size before running the figure.
A dataset that has changed its methodology or definitions mid-series without clear annotation — apparent trends may simply reflect a change in how data was collected.
A chart from a political organisation with a truncated y-axis that makes a small difference look large.
An FOI response that provides data as an image (screenshot of a table) rather than a structured file — request a machine-readable format under the Environmental Information Regulations or ask for a re-issue.
A dataset that covers England only but is described as covering “the UK” in the accompanying press release.

Primary Sources

Office for National Statistics (ONS) — primary statistics for England and Wales
National Records of Scotland (NRS) — population and demographic data for Scotland
Northern Ireland Statistics and Research Agency (NISRA)
StatsWales — Welsh Government statistics portal
data.gov.uk — UK government open data portal
Datawrapper — free, no-code chart and map tool
Flourish — interactive and animated visualisations
Google News Initiative: Data Journalism — free online training
UK JournoHub Data Journalism Hub — tutorials, datasets, and story ideas
How to File an FOI Request in the UK

Jurisdiction note: Statistical coverage varies by nation. ONS data covers England and Wales unless explicitly stated otherwise. Scotland has its own census cycle, run by NRS, which does not always align with the England and Wales census year. Northern Ireland data is published by NISRA and occasionally on different time lags to the rest of the UK. Wales uses StatsWales for devolved policy areas. Always check the geographic coverage in the dataset's metadata before using the data in a story.