Recent grantees have used machine learning to reveal the true scope of oil-well abandonment in Texas; hold land banks accountable in Ohio; and map the proliferation of gold mines in the Amazon rainforest.
These projects harnessed machine learning to augment the reporters’ capacity to tackle big data and systemic issues. The reporters combined the use of machine learning with geospatial analysis, satellite imagery, and traditional shoe-leather reporting, among other approaches.
Another interesting characteristic of the projects we have supported so far is that they all involved collaborative work, whether across newsrooms or disciplines. We encourage journalists to seek smart partnerships that can complement their skills and perspectives.
We're seeking compelling data-driven storytelling—based on original and transparent data collection and analysis—that has the potential to shape public discourse and hold the powerful accountable.
As you prepare your application, here are a few questions for you to think about:
- Why are you using machine learning? Have other approaches been deemed insufficient?
- What data do you plan to use? What biases might this data contain, and how would reporters try to mitigate it? What is the size and complexity of the data you are dealing with?
- What is the machine learning model? Has the model been tested on a smaller-scale dataset? How accurate is the model?
- Can the model be used by others? Can the investigation be replicated with the model?
For inspiration, check out the projects we have supported so far:
- "Security for Sale" | Tyler Dukes, Payton Guion, and Gordon Rago. Access the data, tutorials, and resources shared by the reporters.
- “Waves of Abandonment” | Clayton Aldern, Christopher Collins, and Naveena Sadasivam (2021 University of Florida Award for Investigative Data Journalism, Small/Medium Newsroom). Here’s how the reporters leveraged machine learning to tell the story.
- Properties on the Precipice | Emily Crebs and Lucia Walinchus. Also, read the project’s methodology explainer.
- “The Illegal Runways that are Swarming the Venezuelan Jungle” | Joseph Poliszuk, María de los Ángeles Ramírez, and María Antonieta Segovia
- “Falling Through the Cracks: Pandemic Exacts Heavy Financial Toll on Georgians” | Emily Merwin DiRico, Nick Thieme, and J. Scott Trubey. Here’s a detailed explanation of the data methodology.
In this grant we encourage radical transparency and sharing of methodologies and code, including ethics considerations that informed the research and reporting as well as data limitations, so each story produced can serve as a blueprint for other newsrooms pursuing similar projects.
FREQUENTLY ASKED QUESTIONS
Who is eligible to apply?
This opportunity is open to U.S. residents and journalists around the world. We are open to proposals from freelance data journalists, staff journalists, or groups of newsrooms working in collaboration on a data project idea. We want to make sure that people from many backgrounds and perspectives are empowered to produce data journalism. We strongly encourage proposals from journalists and newsrooms who represent a broad array of social, racial, ethnic, underrepresented groups, and economic backgrounds.
When will you be notifying applicants on whether they've been selected?
We review applications on a rolling basis, as soon as they are received, and typically notify applicants within a month if they're being considered for support. If there is some urgency to the field reporting, the applicant should state the reason in the application.
What is the budget range for the data journalism proposals?
We do not have a budget range for these Machine Learning Grants. We will consider projects of any scope and size and we are open to supporting multiple projects each year. Most awards for our past data journalism project support has been between $10,000-$25,000, but may be more or less depending on circumstances.
Do you pay stipends or salaries for freelance journalists?
We expect news organizations to pay journalists for their work, though in some cases, we may consider stipends to cover a reporter's time, if provided in the budget with an explanation. It is OK to include costs of contractors, such as data researchers or data visualization/story designers in your proposal and budget. Please do not include stipends for journalists/team members who are in the employ of newsrooms or are being paid by a publisher.
What are examples of editorial products or project expenses that the Pulitzer Center grants DON’T cover?
- Books (we can support a story that might become part of a book, as long as the story is published independently in a media outlet)
- Feature-length films (we do support short documentaries with ambitious distribution plans)
- Staff salaries
- Equipment purchases (equipment rentals are considered on a case-by-case basis)
- An outlet’s general expenses (for example rent, utilities, insurance)
- Seed money for start-ups
- Routine breaking news and coverage
- Advocacy/marketing campaigns
- Data projects aimed solely at academic research. Data should be developed to enhance/support journalism.
Send a message to [email protected]!