User friendly and accessible open data on racial disparities
When Zamila Bunglawala, deputy director of strategy and insight at the Race Disparity Unit at the Cabinet Office, and her team of data, digital and policy experts were tasked with creating a website to publish open government data broken down by ethnicity, they had to adhere to two rules.
They had to adhere to “all data and digital standards” in order to uphold and command trust. They also had to communicate with and test the website with several groups of possible users including NGOs, academics, think tanks, local government, central government and members of the public. “That has allowed us to engage with people in a way that is quite new to government, and user testing tells you a lot about policy as well,” said Bunglawala.
Testing the website, Ethnicity Facts and Figures, with potential users proved invaluable. Bunglawala and her team terms found phrases that one might assume are commonly understood turned out not to be. She said that during testing most people didn’t understand how year group or school grade correspond with age, so they decided to just use age on its own. She said: “Giving people ages is quite simple.”
This is one example of testing leading to a better experience for the end user through the elimination of jargon from the site.
“This contains all government data in one place. In order to put it forward we had to make sure there is no jargon, it is user friendly and there were no labels that people didn't like,” said Bunglawala.
The data covers eight topics: population, education, crime, health, housing, work and pay, and workforce and business.
For the site to be transparent and therefore engender trust from the user, the metadata at the top lets them know which department published that data, the date of publication and of the latest update, the source, the geographical area covered and the time period.
Clicking on a topic leads to a list of sub-topic dropdowns leading to pages with summaries. Below the summaries are charts and links to download the source data in CSV format. Each chart is followed by another summary particular to that piece of data.
The page summary also includes three dropdown links; things you need to know, what the data measures and the ethnic categories used in this data. Bunglawala explained that those three points are specifically for academics. She said: “This is the technical information that most academics wanted in terms of legal issues, categorisation issues, etc. Your average user didn't want to see this material but it had to be the site for it to be open data.”
Charts and tables are the only forms of data visualisation on the site. This is because testing had revealed that geographic maps and heat maps were ineffective as most people couldn’t decipher them.
Bunglawala and her team also received feedback that they should not hide any disparities, not try to explain them away and not to give any analysis.
“We have the main facts and figures. On the site so far, you only have charts and commentary. There is no causal analysis of any kind,” she said. “We also had to ensure that there was no judgement or spin on the site.”
Furthering the spirit of transparency, in the instances where data is withheld, the reason is given as to why. For example, a small sample size which makes the data unreliable.
Ever more data is being released highlighting worrying differences between majority and protected groups. At the end of 2018, the Resolution Foundation released research that found that people from black and ethnic minority backgrounds face a pay penalty in the workplace. The most concerning finding is that black male graduates earn up to 17% less than their white counterparts.
The open data made easily accessible on Ethnicity Facts and Figures makes it easier for ordinary members of the public to delve into this issue and gain a better understanding of this and other topics on a national scale.
Zamila Bunglawala was speaking at the Open Data Institute London.