The three kinds of code you write in the newsroom
Journalists have been writing code, in one form or another, for several decades. Yet coding is still considered a speciality skill in the journalism profession.
Thus, if you're a journalist who codes, you probably work alone or on a small team of colleagues with similar technical skills. This divide can breed tension between coders and non-coders in the newsroom, especially when it comes to the rhythms and speed of the news production process.
Journalist and developer Chris Amico knows this struggle all too well. Drawing on a decade of experience slinging code in newsrooms big and small, he tackled this topic in a lightning talk at the National Institute for Computer-Assisted Reporting conference (aka, NICAR) in early March.
“Moving at different speeds is something I've seen every newsroom I've worked in struggle with,” Amico said.
In his six-minute talk, Amico outlined how technical work, like data analysis and web development, should sync with three general newsroom objectives: reporting, storytelling or product.
Amico's remarks were pre-recorded and delivered in absentia. At the last minute, he cancelled plans to travel to New Orleans, La., for NICAR due to rising concerns about exposure to Covid-19. In retrospect, this was a smart move.
Amico also published the script of his talk.
Amico's definition of product is particularly striking. Note how he recasts technical debt and on-going maintenance costs—topics everyone prefers to avoid—as the central underpinning of the newsroom's enterprise.
The NICAR conference is best known for the dozens of sessions that introduce journalists to computer languages, software packages and other technical tools.
For the fourth consecutive year, I co-led First Python Notebook. This year, Andrea Suozzo of Seven Days in Burlington, Vt., joined our coaching team and re-vamped our data viz course materials.
I wasn't just at NICAR to teach. I was a student in First Graphics App, I learned more about how news app developers produce custom interactive content using Node.js.
Here are a few highlights from this year.
Big Local News is data sharing and collaboration platform that launched just in time to demo at NICAR and, as fate would have it, to help journalists cover the Covid-19 crisis.
The Big Local News initiative is led by Stanford computational journalist Cheryl Phillips. Based on previous success with The Stanford Open Policing Project, Phillips believes that, by gathering together disparate data from smaller jurisdictions, journalists and social scientists can identify larger regional and national patterns that yield more impactful research and journalism.
Datasette is a growing ecosystem for exploring and publishing data that fits neatly into the workflows of computer-assisted reporters. Datasette started as a hobby project of Simon Willison (co-creator of the Python's Django framework), who is now working on the project exclusively as a Stanford John S. Knight fellow at Stanford. Check out Willison's tipsheet.
DataShare is a secure document research platform that has become the tool of choice of the International Consortium of Investigative Journalism (ICIJ) when managing projects based on big document leaks. ICIJ researcher editor Margot Williams and Quartz machine-learning journalist Jeremy Merrill explained how Datashare was instrumental in ICIJ's recent Luanda Leaks project.
VisiData is a slick command-line tool for exploring, summarizing and analyzing spreadsheets. It's kind of like a hacker version of Microsoft Excel. Not only is VisiData free and open-source, it's also faster and can handle much larger datasets. Buzzfeed's Jeremy Singer-Vine, who publishes the Data is Plural newsletter, lead a hands-on introduction to VisiData.