People
Who staffs a data team, and what they cost.
Four kinds of people: researcher-reporters, visual journalists, full-stack devs, and data editors.
- Researcher-reporters build datasets, dig into them, and report what they find. They’re comfortable with some statistics, maybe some machine learning, and they work in R or Python about as easily as in a spreadsheet. Other shops call them data reporters, computer-assisted reporters, or journalist-engineers.
- Visual journalists turn data into something a reader can see and use. They have a design eye and usually some sense of interaction, and they live in tools like D3, Datawrapper, Flourish, QGIS,
ggplot2,matplotlib, and Illustrator. Not every data story is visual, but this is where the visual ones get made. - Full-stack devs get it on the page and keep it there. They build the news apps and page generators, tell you what your CMS can and can’t do, and worry about where the thing is hosted. Expect fluency in HTML, CSS, and JavaScript, a framework or two, and some cloud.
- Data editors know enough of all of the above to be dangerous and, more to the point, can run people and ship projects. A data editor doesn’t have to be the best coder in the room. They have to keep the room moving.
Those descriptions straddle editorial and product, which is exactly the tension that decides where the team sits. (That’s Structure.) The people come from everywhere: some went to journalism school, plenty didn’t, and plenty never touched a computer-science degree either. The titles don’t travel, so a researcher-reporter here is a data reporter there and a journalist-engineer somewhere else. What matters is being able to say plainly what a given role does.
Finding them
Data journalism is a small world, and it hires like one. The reliable channels are IRE and NICAR (NICAR 2026 is in Indianapolis), the News Nerdery and Journalists of Color Slacks, affinity groups like NAHJ, NABJ, and NLGJA, and plain job posts. Your best recruiter is usually someone already on the team. As one editor put it, “Our former fellows are our best recruiters.”
You can teach almost anyone Python. You cannot teach reporting instinct or lived experience. When you have to choose, choose the instinct and backfill the code. Pairing a data reporter with a beat reporter on a real project is the cheapest training you will ever run, and it builds the muscle on both sides. As ProPublica’s Scott Klein put it in 2016, recruit people who already have two of code, design, and journalism, and teach them the third. Code is the easiest of the three to teach.
What to pay them
Pay tracks role and, more than anything, geography. The most useful current benchmark is OpenNews’s 2022 News Nerd Survey: non-managerial journalist-coders cluster around $90,000 to $105,000, managers around $135,000 to $150,000, with New York and DC pulling numbers up and most everywhere else sitting lower. Unions and product or engineering roles push pay higher still. Treat those numbers as a snapshot; the bands drift, so check the latest survey before you set one.
Nonprofit and small newsrooms usually land at the bottom of those ranges, so be straight with yourself, and with candidates, about where you stand. Pay is also now the top reason people leave these jobs, which makes compensation a retention problem, not just a hiring one. If a full salary line is out of reach, a fellowship, a grant-funded seat, or a Report for America placement can get someone in the door. That’s Fundraising’s department.
Hiring and paying them is only the start.