O’REILLY RADAR – By Mac Slocum
Data journalism has rounded an important corner: The discussion is no longer if it should be done, but rather how journalists can find and extract stories from datasets.
Of course, a dedicated focus on the “how” doesn’t guarantee execution. Stories don’t magically float out of spreadsheets, and data rarely arrives in a pristine form. Data journalism — like all journalism — requires a lot of grunt work.
With that in mind, I got in touch with Simon Rogers, editor of The Guardian’s Datablog and a speaker at next week’s Strata Summit, to discuss the nuts and bolts of data journalism. The Guardian has been at the forefront of data-driven storytelling, so its process warrants attention — and perhaps even full-fledged duplication.
Our interview follows.
What’s involved in creating a data-centric story?
Simon Rogers: It’s really 90% perspiration. There’s a whole process to making the data work and getting to a position where you can get stories out of it. It goes like this:
- We locate the data or receive it from a variety of sources — from breaking news stories, government data, journalists’ research and so on.
- We then start looking at what we can do with the data. Do we need to mash it up with another dataset? How can we show changes over time?
- Spreadsheets often have to be seriously tidied up — all those extraneous columns and weirdly merged cells really don’t help. And that’s assuming it’s not a PDF, the worst format for data known to humankind.
- Now we’re getting there. Next up we can actually start to perform the calculations that will tell us if there’s a story or not.
- At the end of that process is the output. Will it be a story or a graphic or a visualisation? What tools will we use?
We’ve actually produced a graphic (of how we make graphics) that shows the process we go through:
Partial screenshot of “Data journalism broken down.” Click to see the full graphic.
What is the most common mistake data journalists make?
Simon Rogers: There’s a tendency to spend months fiddling around [Read more…]