Story Discovery At Scale

March 23-24, 2023 at McClatchy Hall, Stanford University

The Story Discovery at Scale conference is bringing together an emerging cohort of thought leaders and practitioners in the field of computation and journalism to address challenges and explore potential collaborative work.

This initial conclave – hosted by Big Local News at Stanford University, and sponsored by The Knight Foundation – is designed to foster conversation, and the sharing of new ventures.

As part of the conference, you will have opportunities to share details on your projects and areas of interest, and we’ll have researchers from Stanford, Berkeley, California Polytechnic State University and Google presenting some of the latest computational techniques relevant to story discovery for the news. This will include experts in specific computational fields who will describe the latest in research and highlighting upcoming developments that can support journalistic efforts.

We hope that this gathering will enable us to avoid duplicated effort, find opportunities to collaborate and scale journalism in support of local news.

We look forward to seeing you soon.

Schedule

Day One – Thursday, March 23

9:00 – 10:00 a.m.

Welcome and introductions

10:00 – 11:00 a.m.

Project Share – In flight or finished

Attendees will have 5 minutes in a lighting-talk style session to present their own projects, with an eye toward surfacing potential areas for collaboration. The focus here will be on projects that are well under way or are complete.

Speakers

David Eads, The Marshall Project – Banned Books in Prisons
Boyoung Lim, The Pulitzer Center – Machine Learning Reporting Grants from the AI Accountability Network
Amanda Hickman – MuckRock/DocumentCloud, the latest greatest features.
Joe Nudell, Stanford Computational Policy Lab – USPS, an address sampling tool big Big Local News and The Markup
Derek Willis, University of Maryland – Open Elections
Simon Willison – Datasette, the latest greatest features

11:10 – Noon

New efforts to track the local and state legislative process for journalistic + public benefit

A variety of projects are helping reporters, citizen journalists and the general public keep better tabs on local and state government. This work ranges from training citizen journalists to attend public meetings to automating the collection of local council meeting documents and videos. These efforts also aim to leverage these raw sources of information in platforms, tools and storytelling. This panel will showcase some of these projects, and provide time for attendees to offer feedback and share tips on similar efforts that may not yet be on our collective radar.

Moderator

Serdar Tumgoren, Big Local News at Stanford University

Panelists

Serdar Tumgoren, Big Local News at Stanford University – Agenda Watch, a local news detection site
Eva Maxfield Brown and Nicholas Weber, University of Washington – Council Data, an open-source project on local council actions
David Lesher, CalMatters, and Foaad Khosmood, Cal Poly University – Digital Democracy, a legislative tracking project of CalMatters
Max Resnick, Documenters – Documenters, a participatory civic media project

Noon – 1:30 p.m.

Lunch

1:30 – 2:30 p.m.

Expert Share – Datacommons.org

Learn more about using and working with Datacommons.org, an effort to provide access to cleaned, normalized and joined open data, supported by Google.

Speaker

Ramanthan Guha, Data Commons/Google

2:30 – 3:20 p.m.

Finding and addressing the friction points

Is there a project challenge you’ve faced that seems insurmountable? Was the obstacle technical, financial or both? We’d like folks to share these challenges (and any possible solutions you’ve found) during this informal panel.

Some examples of friction points:

Document processing at scale (e.g. ingestion/indexing)
Audio/Video processing
Packaging and deploying tools and products to newsrooms
Removing boilerplate from documents (better search, NER, etc)

Moderator

Ben Welsh, Reuters

Panelists

Dustin Dwyer, WUOM-FM at University of Michigan and the Minutes project
Aditya Parameswaran, University of California, Berkeley and the Community Law Enforcement Accountability Network (CLEAN)
Lisa Pickoff-White, KQED and California Reporting Project and CLEAN
Heather Bryant, Tiny News Collective

3:30 – 4:30 p.m.

Topical breakout discussions

4:30 – 6:00 p.m.

Break

Take some time to explore campus.

6:00 – 9:00 p.m.

Dinner

Day Two – Friday, March 24

9:00 – 10:00 a.m.

Project Share – Up and coming

Another lightning-talk style session, where presenters will have 5 minutes to share early-stage ideas or projects, what they need to make progress, and how best to engage with them.

Speakers

Michelle Dragony, Coastside Buzz – A local news information business model
Darla Cameron, The Texas Tribune – Automating school board elections
Lisa Hunter, LION – News Sustainability Audits + automations
Brant Houston, Knight Chair, University of Illinois – RiskMinr, a one-stop data shop for data and info on big agriculture companies
Jerry Zremski, University of Maryland – Maryland Commons

10:00 – 10:50 a.m.

Expert Share – What's next in news detection and extracting structured data from documents.

What might a computational newsroom look like: one in which journalists are assisted in the process of finding, sourcing, structuring, and updating stories. Every story is unique, but in our research, we show that there are common patterns to the decisions journalists make — why certain events are written about; why specific sources are used; how background and historical context supplement main events in writing; how breaking news updates unfold. In our work, we’ve trained models (e.g. GPT3, and others) to learn these patterns and are working to operationalize them within Big Local News. We’ll discuss the concrete advances we’ve made, the steps that need to be taken, and the pitfalls and limitations we need to be aware of.

Speaker

Alexander Spangher, University of Southern California

11:00 – Noon

From data streams to news

Are you gathering data on a regular basis and feeding it into some aspect of your content creation pipeline? For example, Big Local News produces alerts for reporters about the latest federal layoff notices, campaign finance filings, and federal audits. We want to do much more to improve the signal-to-noise ratio in terms of detecting news in these streams and providing digestible outputs to seed news stories and assist the reporting process. On this panel, we discuss what’s happening in this space and solicit feedback and comment from the audience. We’d love to hear what you’re working on in this area, or what you wish was possible (even if you haven’t yet solved the problem).

Moderator

Eric Sagara, Big Local News at Stanford University

Panelists

Darla Cameron, The Texas Tribune
Dan Kopf, San Francisco Chronicle
Silas Lyons, Gannett/USA TODAY Network’s Center for Community Journalism
Eric Sagara, Big Local News at Stanford University

Noon – 1:30 p.m.

Lunch

1:30 – 2:30 p.m.

A Blue Sky/What If: Conversation on the present + future of AI-driven storytelling

Panelists will share real-world applications of AI they’ve encountered in newsrooms big and small, along with their thoughts on areas of future promise (and need) for journalism both about and with AI.