One of London’s poorest boroughs is showing the rest how to do data

When Pye Nyunt was recruited by Barking and Dagenham council in 2016 to lead its new Insight Hub, the borough lay at the bottom of most of London’s leaderboards.

In 2015, its unemployment rate was the highest in London and its educational attainment was the lowest. Its long-term business survival rate and its income were lower than the London average and that of its neighbours, and it had the twelfth highest deprivation rate in the country.

The Insight Hub, a team of data and behavioural scientists, was tasked with, among other things, finding new ways to deal with the borough’s deprivation and social problems using digital techniques.

“In 2017, in the London Borough of Barking and Dagenham, we conducted our largest public conversation to date. We asked our residents what they like about their borough, what can be improved and what their aspirations for the borough were” said Nyunt. “We had over 3,000 responses, which we pulled together in a community document called the Borough Manifesto. The data was collected into key themes and targets that we have set for ourselves.”

“What I believe we are doing is democratising data”

To achieve these goals, which are to be reached inside the next 20 years, and to track the borough’s progress towards them for all to see, the Insight Hub is now trialling a brand new tool. In partnership with Future Cities Catapult (FCC), a London-based innovation lab for cities, the borough is using a digital connector to pull together and merge different data sets from the eclectic sources which make up the manifesto’s targets. Once together, they will be used to build an index, to track the council’s progress against the people’s aims.

“We said, wouldn’t it be great if we could, via the digital connector, pull together all of the different datasets that we’ve put into the Borough Manifesto, and create a digital index with a visualisation, so we can show our residents how we are actually doing against the targets that have been set,” said Nyunt.

Plenty of cities and councils in the UK have open data portals, where the public can download data in a raw form. Barking and Dagenham aims to use the digital connector to go a step further.

“It’s not solely designed to be an open data platform – it’s taking the conversation we’ve had with residents, but then plugging that data into something that automatically updates and visualises itself, so that residents can understand how the borough is doing compared to our neighbours and the rest of London,” said Nyunt. “Doing this helps us hold each other to account: what I believe we are doing is democratising data.”

Unlocking the data science potential in British cities

There’s huge potential for data science to make government work better. By analysing masses of interconnecting information, data scientists can find the patterns that can help government speed up and improve many of its core services. Several cities in the US have had big successes and in Australia, Queensland revolutionised the way its hospitals work with by using data science to predict and prepare for surges in demand.

“It should be easier to share data than it was two and a half years ago”

However, for local government in the UK, things have been slower to take off. Despite a recent index putting the UK government on top of a list of OECD countries as the most prepared to begin working with artificial intelligence, cities and towns in the UK are yet to really get to grips with its potential, with the exception of small experiments run in individual departments of central government.

Part of this is because of how difficult it is to get the data together. In London’s most ambitious data science project to date, the first of its newly-formed London Office for Data Analytics (LODA), which amassed and analysed multiple sets of data from across 12 London boroughs, data scientists were forced to spend three to four months at the beginning of the project cleaning and linking the data sets into one set of information they could work with.

“We wanted to build an open tool for the public sector to do things with data”

“It should be easier to share data and/or models than it was two and a half years ago,” said Dr Joseph Bailey, the Data Science Team Lead at FCC.

“We wanted to build an open tool for the public sector to do things with data by focusing on a really quite niche problem,” said Jon Robertson, a project manager for Tombolo. “Focusing more on getting the data into the right shape, connecting the data and getting a common output that could be shared.”

The digital connector pulls together data from across the city allowing it to be analysed in combination

The tool is designed for data science specialists to help make the painstaking job of linking datasets quicker and easier. When data comes from different sources and covers different factors and geographies, getting it into the condition where it can be analysed takes time. The Tombolo connector is designed to allow civil servants to apply data to real problems much quicker.

“It can be quite challenging to combine data from different sources, because they’re in different geospatial granularity and there’s a lot of inconsistencies in how it’s provided,” said Bailey. “We’ve created something that pulls in all these different kinds of data, normalises everything internally, so that everything’s on a consistent database schema, and then enables you to share that as you choose.”

“What we’ve found in a lot of councils is that the data people are typically not given that free rein to do data projects”

The Tombolo digital connector is built using open source software and is free to download for any organisation. The project was funded by Innovate UK, and the FCC team hope to conduct further trials of the tool with other cities and local authorities in the UK in order to refine the tool and test its applicability.

Local government in the UK is often offered software solutions that make use of data, but the sources of data they use can become quickly outdated. The digital connector allows local government to use combine the data sets they need, be it open data or their own private information, onto one server to do the analysis they want.

“What we’re doing here with FCC is building a tool that we know that local authorities can make best use of because it’s bespoke and it’s designed for them in mind,” said Nyunt. “It’s not a software solution that we need to purchase. We’ve been involved in its design, its build and its testing so we’re confident we’ll have something that is fit for purpose.”

A data science skills shortage

Whether the tool does, in fact, get used widely local government in the UK however rests on another issue – whether or not they have the data specialists to take advantage of it.

On Barking and Dagenham, Robertson said: “They’re one of the only councils in the UK that have gone out and recruited data scientists and get them in-house. What we’ve found in a lot of councils is that the data people are typically doing monitoring and reporting, and they’re not given that free rein to do data projects.”

Even within boroughs in central London, applying data science to policy has been slow to take off. Lack of staff, and misapplication of their skills towards reporting or updating dashboards, is part of the problem.

“There still needs to be more investment from local authorities in getting people in who know how to do data science”

“We’re a city of 8.6 million residents, yet if you think about London’s councils, I reckon there are around 10-15 data scientists, or 10-15 individuals who you could realistically attach that label to,” said Andrew Collinge, Assistant Director and Smart City Lead at the GLA.

Local government in the UK is in a double bind. It has difficulty competing with the private sector to recruit the data scientists it needs for ambitious projects, and as a result, it lacks the expertise to attempt the ambitious and interesting projects that might attract them.

“Right at the time they most need to find different ways of doing things, they have the least resources available to do that,” said Eddie Copeland, Director of Government Innovation at Nesta. “If you’re good at data science, and you live anywhere near a major city, there are private sector organisations that will pay you a lot more than the public sector is able to.”

“In the time that we’ve been working on the project, we expected that the public sector would push on and start to recruit and get more advanced data skills, but there’s still a bit of a lag,” said Robertson. “There still needs to be more investment from local authorities in getting people in who know how to do data science.”

(Picture credit: Flickr/Nico Hogg/Future Cities Catapult)