How data-driven storytelling project Crosstown helps scale neighborhood news in Los Angeles


Local Media Association is profiling some of the first Google News Initiative Innovation Challenge winners in North America and how they’re putting funding to work developing sustainable business models, diversifying revenue streams, and increasing audience engagement. Watch the webinar from Sept. 29, which explores lessons from hyperlocal news expansions, including Wick Communications, Torstar Local, and Crosstown — all leveraging new technology to connect with untapped audiences.

By Joe LananeLMA Contributor

The demand for hyperlocal journalism is higher than ever, but so are the costs to produce quality reporting at the community level. One academic-driven project recently received financial support from the Google News Initiative Innovation Challenge to efficiently scale neighborhood news.

Crosstown, a Los Angeles-focused data-driven hyperlocal news site, is run by the University of Southern California’s Annenberg School of Journalism and the Viterbi School of Engineering. Professor Gabriel Kahn doubles as the editor/publisher of the online operation, collaborating with student journalists and the college’s computer science department to generate neighborhood-level data across Los Angeles and L.A. County.

The project started in 2018 and is on the cusp of a milestone. The Crosstown Neighborhood Newsletter will soft launch in October with custom editions for each of Los Angeles’ 110 neighborhoods. The hyperlocal content — data visualizations and story text — is auto-generated to be unique to each area using city-provided data mapped and organized per community. The model incurs no marginal cost for producing each unique newsletter product, which Kahn said will eventually be used to attract paying members as well as advertisers interested in geo-targeting their hyperlocal audience.

“L.A. is a media capital but suffers from news deserts as much as rural areas,” Kahn said. “Our solution takes advantage of the data that so many large metros have started to provide to the public.”

Crosstown maps can help L.A. residents learn the specific crime rate in their neighborhood, among other data sets.

Instead of declaring whether crime is up or down citywide, for example, the project can help L.A. residents learn the specific crime rate in their neighborhood.

“Those city numbers just don’t correspond with residents’ lived experiences,” Kahn said. “The more you can localize data, the more you can write stories that relate to people’s lives.”

Crime data is one of several sets that can be analyzed at the neighborhood level for potential stories. Such drilled-down information usually is expensive to produce and report about. The CrowssTown project has streamlined such efforts by pulling data from Los Angeles Police Department and the L.A. Sheriff’s Department. Kahn calls it an evolution of the police beat, delivering crime information with visual context rather than reporting on crime-based incidents.

As part of GNI’s North America Innovation Challenge, the project received $270,000 to develop the neighborhood-level newsletters that could eventually lead to a membership model. The newsletter will include a big-picture perspective breaking down the findings of each dataset — up to three different topics delivered weekly — along with neighborhood-specific analysis that provides for visual context comparing the numbers to other areas of town over time.

Data sets cover a slew of quality-of-life issues, ranging from the number of potholes filled to how many pool permits were issued during the pandemic. Many of the projects have driven the L.A. news narrative in recent months.

For example, Crosstown produced a statistical breakdown of the most dangerous intersections in Los Angeles using traffic accident data, drawing significant coverage from other media. The project also framed car burglaries across the city, and an effort is underway to explore air quality per region in L.A.

Kahn said the project is only possible because Crosstown collaborated with USC’s computer science students and faculty, who help capture data sets in visualizations, as well as full paragraphs of custom text per neighborhood. The tech experts are also able to track breaking news and statistical anomalies that tip off reporters to a story sooner than usual.

“This is an exciting application of our artificial intelligence and data science research at IMSC, as we need to acquire, manage, analyze and visualize a large amount of citywide data in real-time for journalists’ use so they can see the patterns that would translate to news stories,” said Cyrus Shahabi, chair of the USC Department of Computer Science and director of Integrated Media Systems Center, in a statement.

Kahn compares the project to similar efforts done at Outlier Media, which distributes statistical breakdowns of Detroit’s housing and utility news to underserved communities via text message.

“Sarah [Alvarez of Outlier Media] pioneered this,” Kahn said. “I think we can be that multiform ‘Outlier’ in a way by telling people all sorts of things about their neighborhood.”

The goal is to make Crosstown a viable news source for Los Angelenos, according to Kahn, but ultimately he hopes the effort scales across other newsrooms.

“I think it can be replicated in any mid-size to large metro area and then also possibly at the state level where you’re collecting county-level data,” Kahn said.