In the interest of being part of the solution, I thought I’d interview a few of my colleagues here at Atlantic Media who had graduated recently from journalism school to see what they wish they had learned, if not more code.* None of the colleagues I spoke with is a traditional print reporter, but among us, we do the types of writing, editing, and data analysis that many reporterly types aspire to. Here’s what we wish we had learned before becoming journalists:
It’s remarkable the number of times I’ve wished I knew how to perform a regression analysis. I took some economics and statistics in college, but it was barely mentioned in j-school, even though most journalists have to analyze numbers and determine whether and how to report on them.
Jordan Weissmann, who covers business, agreed:
Really, all stats. [My professors] never really communicated— and honestly, I don't think really understood -- how crucial a firm grasp on math would be to succeeding in certain corners of the business today. You don't necessarily need to be able to do all the math—that's what Excel's for -- but you do need to understand what it basically means. That's far more important than doing man-on-the-street interviews.
And it might even be worth it if you’re planning to cover entertainment. Ashley Fetters told me:
I have an uncle who's pretty sure all journalists should be required to take macroeconomics—so that all these 'bleeding-heart liberal journalists' who just think every cause deserves more money understand where money has to come from. Which is ... on one level, somewhat valid.
Data analysis is one of the best ways to find story ideas, especially if you are, like many student reporters, not very well sourced. The course syllabus in the most basic online class in my grad program has changed—I think for the better—since I left. There’s now a data component that includes a section on Google Fusion Tables, which I taught myself in grad school and still use occasionally. (I file this under “not coding,” because it involves plugging a spreadsheet into an existing application, which spits out the embeddable code for you.)
The most valid criticism of my article was that journalists need programming skills to find and scrape data for stories. I did my grad-school data reporting with Excel, but I would actually love to be better acquainted with database platforms like MySQL (which is not mentioned in the more recent syllabus). I am not sure, however, whether MySQL queries are considered “coding.” Is writing formulas in Excel coding?
Easy-to-use online tools designed for non-programmers, such as import.io and Google Scraper, can help you do the scraping necessary for most data projects. But it’s true that messier or harder-to-reach information might require programming languages such as Ruby and Python.
David Yanofsky, a reporter who specializes in data visualization at our sister site, Quartz, explained how he uses data scraping to create stories. He went to design—not journalism—school, but he has good insights for those who want to take the plunge and learn Python or R:
With Python I can scrape information from websites (that's how the wager data from the NFL betting piece last week came about) as well as analyze and calculate complex forms of data like the center of gravity of the worlds airline routes.
Effectively code has become my reliable, repeatable, and powerful calculator.
This is especially the case with how I use the language known as R. I used R for every calculation in my story about baseball ticket prices. A couple of times in the process of writing that story I found or the provider of the data found inaccuracies. Since all of my calculations were in script form, when new data came all of the numbers I calculated before could be updated in a matter of seconds.
If I had one or two years of grad school to become an effective reporter, though, I would try to save time by making full use of the ready-made scraping tools geared toward non-programmers, unless I really needed the programming skills for my specific project or story. (In this case, the “everyone needs a little bit of code” argument is flawed, too, since you can’t “little bit” scrape some data.) And in a professional newsroom, I would still leave the creation of the graphic about that data to someone with design training.
It's a pretty safe bet that on any given day, TheAtlantic.com will have a story featuring at least one study. Several of us agreed it would be much easier to evaluate their validity if we had taken a crash-course in experiment design. How big is a big-enough sample size? Which journals have a strong peer-review process? Which findings are significant, and which are garbage?
Several people brought up the delicate art of pitching freelance stories. Most writers do at least a little freelancing when they’re starting out, but other than "push send and pray," there's usually little strategic training in that realm.