The Race Disparity Unit, Cabinet Office, have been announced as a finalist in the DL100 Digital Team of the Year category for 2018. The Race Disparity Unit was formed from 3, starkly different disciplines: Data Analysis, Policy, and Digital. Against a backdrop of intense scrutiny, tight timescales and huge cultural differences, this team came together, delivering The Ethnicity Facts and Figures website. This is truly democratised data – making 200+ previously hidden datasets publicly available and transparent.
On a summer day in 2016, Theresa May stood on the steps of 10 Downing Street and called for an audit of all government data on ethnicity. In her words, the purpose of the audit was to “identify and publish information showing how outcomes differ for people of different backgrounds [..] to reveal racial disparities and help end the burning injustices many people experience across Britain.”
Shortly after the Prime Minister’s announcement, the Race Disparity unit was established in the Cabinet Office to carry out the audit. Methods were selected to join the unit at the end of discovery to deliver the website behind the audit. Looking at the complete Ethnicity Facts and Figures website you’ll see that whilst it’s dense on statistics, it’s all displayed in plain English. The website launched on October the 10th with more than 100 datasets that show how different ethnicities have different outcomes across topics such as education, work and crime. It is the first time that an audit of this kind of data from across departments has been collected and presented in a uniform and simple way.
We quickly identified that the challenge was to consolidate data from a wide range of disparate sources and present it in a website that could be easily interrogated and would be accessible and meaningful to different users. At Methods, we put users at the heart of everything we do. Our researchers established a spectrum of user types, from members of the public with low data literacy, to expert users who work with ethnicity data. Across the spectrum of users, we identified very different user needs. We continually did usability testing with all user types to make sure that our service would meet their different needs.
The presentation of the data in the form of graphs, tables, commentary and raw data downloads was developed based on insights from the continuous engagement and testing with users. We tested many different options and iterations, including maps and other more interactive tools, but we found that the data needed to be presented as simply and clearly as possible leading to the current design, which tested well with all user groups. Further, this way of presenting the data has been recognised as an “innovative way of presenting statistics [..] that shows how Government departments could do much more to publish data with the public users in mind – rather than simply publishing data in the way they always have done.”(Ed Humpherson, Head of the UK Statistics Authority).
It took a lot of work from the data analysts and our data experts to get the data in the uniform style we needed. On Data.gov.uk you’ll notice that all departments generally do data publication their own way, with various surveys and publications being formatted differently – so we weren’t surprised to discover early in the project that departments have their own unique way of measuring and reporting ethnicity. We needed consistency and with help from our data strategists, we developed a data template, which successfully brought together datasets from across government departments and sub-departments as a common data resource.
Our work and the website (www.ethnicity-facts-figures.service.gov.uk) itself has received a lot of positive feedback and has been recognised by industry experts as a “substantial achievement” (Ed Humpherson), an “extraordinary website” (The Washington Post) and a “clear and user-friendly” website (Damian Green, First Secretary of State). Looking back, we ascribe this achievement to our user focused cadence of development which always revolved around testing, learning and iterating.
This article was originally published here.