The Bureau Local, a Google-funded investigative unit headquartered in London, had only been going for a month when Prime Minister Theresa May called a snap election in April 2017, turning any plans its new director had on their head.
Megan Lucero had been hired to the Bureau Local by its parent organization, the nonprofit Bureau of Investigative Journalism, in January 2017 to build a network of UK journalists to collaborate on data investigations. Lucero, a California-native, had spent seven years at The Times of London, launching its data journalism department from scratch. Her first job at the Bureau was to get out of London and on the road, finding out what work data journalists were doing in the corners of the UK. On the way, she gathered a mailing list of local reporters interested in the project.
Then on April 19, journalists hit panic stations. Many of them—in shrinking newsrooms, working for papers syndicated across several cities—said they had their work cut out for them just keeping up with covering the upcoming election.
The election was a chance for the Bureau Local to establish itself as a resource for overstretched reporters.
“Everyone we talked to in our network said they had no time to investigate anything,” Lucero remembers. “I had barely gathered a team—we had only been together a month.”
Rather than fight the turning tide, Lucero and her team recognized that the election was a chance for the Bureau Local to establish itself as a resource for overstretched reporters. The plan worked: The network grew to 350 regional reporters spanning 84 cities across the UK over the following weeks. It produced a total of 77 stories during the election, reporting on political advertising on Facebook, reporter access to party leaders on the campaign trail, and new voter registrations, accurately predicting where seats were at risk of changing hands.
The Bureau Local was founded as part of the Bureau of Investigative Journalism in London using a grant from Google. Now in its fourth round, Google’s Digital News Initiative has awarded €70 million ($82.5 million) to more than 350 journalism projects in 29 European countries. The Bureau Local’s grant, worth €662,000 ($780,769), was awarded in February 2016 for a term of three years.
Just a few months after the grant was announced, the shock results of a referendum on the UK’s membership of the EU demonstrated the weakness of existing local reporting. Almost no one saw Brexit coming.
Local news was failing to fulfill its democratic function. Two thirds of local UK districts and 56 percent of the population are no longer served by a local paper, according to a 2016 study by Kings College London. Just four publishers account for 73 percent of local media titles. Kings College identified an urgent need for further research to inform potential policy interventions. But neither policy interventions nor extra funding were forthcoming. So the Bureau took a different route. “This whole project was about flipping everything we know about journalism on its head,” Lucero says. “Everything was new.”
Rather than bolstering the competitiveness of existing newsrooms, the Bureau Local wanted to test a different approach to holding those in power accountable using a distributed network of reporters, data scientists, and community activists collaborating on data. “It is really important to us that we’re not propping up media outlets,” Lucero says. “Our strategy has been playing to strengths of the individuals.”
A week before the election, the Bureau put its network to the test with a collaborative hack day. Lucero and the team planned to build the UK’s first digital database on voters by constituency, the UK term for a geographical area represented by a politician. They noted voters’ age, income, and education, hoping these extra details would help them determine the result of the election.
The hack day was planned for four cities across the UK: Birmingham, Bournemouth, Glasgow, Cardiff, and London. Matty Edwards was a journalism student when he learned the London hackathon would be held at his university, Goldsmiths. On the day, attendees worked over the constituency data methodically, with some looking into individual constituencies and others pursuing national angles.
Edwards was looking into a London constituency called Croydon Central. It also happened to be one of the most marginal seats in the country, where the Conservative Party, the center-right party in the UK, was at risk of losing power. He noticed that the Bureau Local estimate for newly registered voters there was higher than the margin the Conservatives had won by in 2015.
Edwards’s discovery became the basis for a national investigation into how new voters could dent the prime minister’s election majority. The Bureau Local identified 71 constituencies where there were larger numbers of new voters than the majority. The modeling led to a national story in The Times of London and 18 local pieces. On election day, nearly half of the seats identified did indeed swing.
“It felt great to work alongside people on the story, even though we had only met that day,” Edwards says. “I ended up getting an additional reporting credit on the Bureau’s website, which was an important step for my portfolio, and a version of that piece was published in The Times.”
After the election, the Bureau appealed to its network for datasets that could be relevant across the UK. In Bristol, the founder of a media cooperative thought he had something. Adam Cantwell-Corn had used multiple freedom of information requests to obtain data from the government about immigration raids in Bristol and ten other cities in the UK. It took him eight months to acquire data that showed that in every city, the overwhelming majority of times a citizen was stopped, they turned out to be British. Cantwell-Corn spoke to legal experts who all said the same thing: “The upshot was that this was pretty strong evidence of racial profiling.”
The Bureau Local distributed Cantwell-Corn’s data across the network. “We had to work out what that process was and the different expectations towards one another,” Cantwell-Corn says. “It goes against the journalistic tradition of guarding your story. It helped that working in different areas, we’re not competitors for market audience.”
Regional reporters added human faces to the numbers. One reporter spoke to a British citizen living in Wales who said he felt like “collateral damage” in the UK government’s pursuit of illegal migrants when the Chinese take-out shop he worked in was raided during his shift. Another British citizen working at an Indian restaurant in Yorkshire, the North of England, reported that immigration officials had only stopped Asian staff, not white staff, during a raid that had hurt the business.
Stories ran in the Bristol Cable, Yorkshire Evening Post, Liverpool Echo, Overtake, Ferret, Birmingham Eastside, Birmingham Mail and Sheffield Star. The story prompted eight politicians to call for the government to review practices. Not bad for a collaborative project where reporters were not paid outside of their normal jobs.
Cantwell-Corn says the funding question has proved to be a limiting factor: “The problem is that a lot of local journalists are slammed and strung out by the amount of work they have to do, so they aren’t even digging up data to build these stories.”
Data journalism is expensive because it takes time: That’s why it is often one of the first functions to go when budgets get tighter. It is expensive in the UK, where, Lucero says, government data is not readily available and then often needs cleaning before it can be distributed to reporters. “The thing that would make my life easier is if the government published data transparently and consistently,” Lucero says. “We spend half our life cleaning datasets.”
That’s where Lucero’s next plan comes in. The Bureau Local has secured £100,000 ($133,720) in funding from the Open Society Foundation to pay (among other things) for a community coordinator to support local journalists in their work. There will also be some money set aside to pay people for their time on investigations. “It’s not my job to fix the business model, but I can’t not think about it either,” Lucero says.
The Bureau will test paying individuals with some seed money to see if it works, and if it does, Lucero says she will seek out longer-term funding. “People are very passionate and very angry [about local journalism],” she says. “By no means am I going in to tell people how to do their job. We go in and say we’ll stand next to you and do something with you. We have to do that at scale.”
Correction: This story previously misstated the number of cities involved in hack day, and then number of cities investigated for the immigration raids story.