5 takeaways from Mapping Local News: An Automated Approach

Chelsea Cirruzzo, (@chelseacirruzzo), a volunteer with the ONA Resource Team, compiled these key moments from the ONA20 session on Oct. 9, 2020. To view a recording of the session, register for on-demand access to the ONA20 archive. Session participants included:

  • Ana Mendez, Special Projects Editor, Lenfest Local Lab
  • Moderator: Michael Krisch, Deputy Director, Brown Institute for Media Innovation

5 key takeaways:

  1. The mapping prototype created by the Brown Institute and the Lenfest Local Lab can allow local newsrooms to map the locations of their articles, which can be used to identify areas of coverage that may be missing. According to the presenters, the goal of the project is to create an open-source tool for newsrooms to extract and map locations from articles. The Philadelphia Inquirer is already partnering with the two organizations. Any other newsrooms that want to get involved can reach out directly to LenFestLabs (@lenfestlab) and the Brown Institute (@browninstitute).
  2. This work can be challenging, Krisch explained. For example, one article from the Philadelphia Inquirer can contain references to different locations in one single sentence, which can confuse the AI.
  3. When the tool was used for the Philadelphia Inquirer’s Built Environment desk, it showed there was a lot of coverage concentrated in center city Philadelphia, where a lot of development is happening, as well as along the river. The newsroom had already been having discussions about which neighborhoods it was and was not covering and if they weren’t covering a neighborhood, what stories existed there. Initial observations from the newsroom included asking if there was a way to overlay maps on crime rates, median housing prices, race, etc. Ana said the maps can show the breadth of coverage by a newsroom.
  4. There are some questions the developers want to consider as they work on the prototype: What do we do about uncertainty around the tool? What level of uncertainty is acceptable in a production environment? What new products can be imagined to address issues made evident by the tool? Aside from geography, what other data can be overlayed over the tool? From a business perspective, what new products can be built that leverage location information?
  5. “A lot of looking at coverage can feel a little ungrounded until you provide a sort of benchmark to measure it against,” Krisch said, which is why the prototype hopes to bring in other data sets to contextual data to answer questions like, why is one area a hotspot? Will other data explain why?

Memorable/tweetable quotes:

  • “One of the benefits of working within a local newsroom is there are so many idiosyncrasies and so many hyperlocal words that are relevant in a really specific context,” Mendez said. “You really need the lens of a local newsroom to anchor that work.”
  • “Once you do know location, it really points to a lot of new ways that you might be able to meet new audiences at a hyperlocal level,” Krisch said.
  • “There are ways to use this tool to access for equity … in ways that are not geographic specifically,” Mendez said.

Links to additional resources