Digital Innovation Top Stories
5 mins read

Reinvigorating “news deserts”: Data journalism tools for newsrooms

The decline of local journalism in recent years has been alarming. Some communities have become “news deserts” as their local papers have either disappeared or reduced to being empty shells of their former selves. Writing in The Washington Post, Steven Waldman, President and Co-founder, Report for America, and Charles Sennott, CEO of the GroundTruth Project called the “crisis in local journalism, a crisis of democracy.”

“Rocket fuel for local journalism”

Two companies working in the space of data journalism have built solutions to help understaffed local newsrooms scale up the production of data-driven local stories.

The Associated Press has expanded its team of data journalists in the last three years, leading to a substantial increase in the number of data distributions supplied to local newsrooms. It is also collaborating with data.world, a platform for modern data teamwork, to make it easier for local newsrooms to discover and use reliable, localized data that enriches reporting.

The way it works, according to Troy Thibodeaux, Data Team Editor at AP, “Whenever AP produces a large data-driven project with granular data (for example, with data points for every county or for a large number of cities), we package it up with supporting documentation and customizable queries that news organizations can use to localize the story. We deliver the data via the data sharing platform data.world through a special AP organization that includes our data distribution members.”

AP’s data distribution efforts are like rocket fuel for local journalism. In 2018 alone, our data was downloaded nearly 1,400 times by journalists in more than 300 local newsrooms. That translates into hundreds of local stories from across the country that would not otherwise have been told. When paired with AP’s own state- and national-level reporting, this journalism provides a rich, multi-layered look at topics that are essential and newsworthy.  

Brian Carovillano, Managing Editor, Associated Press

Some examples in AP’s press release stating how local news outlets used its data to positive ends include:

  • Housing and Urban Development Secretary Ben Carson withdrew his proposal to raise rent for people who get federal housing subsidies after The Detroit News, with the help of AP data, reported that the proposal would lead to rents rising by 20%.
  • Several school districts, including Broward County, Florida, and Denver, Colorado, decided against seeking grants from the National Rifle Association based on the AP’s report on the number and value of those grants.

Data is often the strongest source for a story – it can help lift a story beyond the anecdotal and provide a common point of reference for differing perspectives,” says Thibodeaux.

However, he adds that “far too few local newsrooms have the skills on staff to make the most of the opportunities data journalism can provide.” That’s why the project expanded to offer support to reporters who didn’t have data skills. For example, for complex data sets, the AP has a webinar that walks its users through the data.

Eugene Tauber, Senior Data Journalist with The Morning Call which has used AP’s datasets says, “In most instances, the AP provided context and expertise on the datasets that we would have difficulty doing on our own. Through the addition of webinars, narratives and added data definitions we were able to localize and contextualize complex data in short order.”

While the AP didn’t charge newsrooms for the use of its datasets last year, this year they plan to do so. The details have not been made available yet.

Scaling up news production

The other company working in this space is RADAR AI. It is combining humans and machine to scale up local news production across the areas of health, education, crime, transport, housing, and environment.

The RADAR project is a collaboration between the UK Press Association and a startup called Urbs Media. It has a team of only five data reporters and two editors that produces about 8,000 local stories per month across the UK. These stories are run by the various local media outlets that subscribe to RADAR’s wire service.

Alan Renwick, Director at RADAR says, “What we actually do is take national datasets and do a fusion of local and national journalism. We try to understand the generics of national data and see how these stories might vary for every local area. Then we look at the different variables to figure out how many different outcomes there might be and write story templates for every eventuality.”

“For instance, we might put the numbers from a spreadsheet into a sentence such as “since w, house prices in x have increased by y/fallen by z/stayed the same.” This means that if you have 500 rules you might have 500 different stories. From the same data, you might have different types of stories with different headlines and different content.

Tim Robinson, Group Content Development Director at JPIMedia, a news publishing chain subscribed to RADAR’s wire service told CJR, “It (RADAR) enables us to cover subjects which we potentially wouldn’t be able to cover at all, or certainly in volume.” JPIMedia publishes 95% of RADAR’s stories as-it-is with only minor changes to the copy and addition of geographically resonant headlines. Robinson says that their websites have seen a boost in the number of page views since they started including content from RADAR.

One of the strengths of RADAR’s system is that it uses human journalists to determine the angles, trends, or outliers that are newsworthy in the data. They structure the template for the various versions of the article accordingly. The automation then helps adapt the writing locally, and if required a local journalist can work on it to improve local relevance.

RADAR’s wire service was initially offered for free with the help of a grant from the Google Digital News Initiative in 2018. But now the company is looking to become sustainable and beginning to sign up paying customers. Its pricing model will be based on how many local geographic areas a publisher wants to cover. The company has not shared other specifics yet.

Here’s a webinar by RADAR AI which gets into the details of how the system works:

“The pillars of self-government”

A recent study by the University of North Carolina’s School of Media and Journalism found that over 1,300 US communities have lost news coverage. The researchers commented, “In an age of fake news and divisive politics, the fate of communities across the country, and of grassroots democracy itself, is linked to the vitality of local journalism.”

Professor Philip M. Napoli of Rutgers University who lead a separate study of three demographically disparate communities in New Jersey commented, “If journalism and access to information are pillars of self-government, these tools of democracy are not being distributed evenly, and that should be cause for concern.”

Telling a story with data often provides a way of conveying a bigger picture, while illustrating local implications by highlighting a localized aspect of a dataset. Data-driven methods can also help aid the watchdog role of journalism, which is especially needed in local journalism.

Kira Schacht, Co-founder, Journocode, a German data journalism start-up

Local newspapers have been cutting down on their workforces due to declining revenues; nearly half of all newsroom jobs have vanished in the last two decades. In such an environment, tools like these may play an important role in ensuring the health of local journalism… and democracy.

Related posts