Martin Belam is editor of Ampp3d, a project by The Mirror that uses data journalism and visualisation to tell stories in a clear, concise and fun way. The site is designed to be viewed on mobile devices, and experimentation with visuals have fuelled some debate online about accuracy. Here he discusses this tension between accuracy and clarity in data driven journalism.
Why did you decide to use straight lines on your map of the UK when representing YouGov survey regions?
I put straight lines on the map because the fidelity of that bit of information just didn’t seem to be important in this story. What I spent a lot of time fiddling about with was the image sizes and resolution, and uploading them so that on the mobile version of Ampp3d the map could be seen in totality in one screen size. That seemed more important to me that fiddling around with map shapes.
My feeling was that there is effectively only 5 data points on the map. In the original article, one of the things I wanted to address was why does Ukip get so much more airtime than the Greens, who have an MP. An element of the story-telling was the density of the colour of the map. If you look at the article, I set the opacity of the Photoshop layer to twice that of the level of support that the party was getting. So Labour’s 40% support somewhere meant a region at 80% red. Ukip’s 17% support meant an area appearing 34% purple. The Green level of support, in single figures, essentially meant their map remained white - and that was the design imperative.
The article had a jokey caveat about the straight lines, and I didn’t think much more about it. It was published about a month before anybody really noticed it on social media and questioned it. Of course, all of that context gets stripped out when the map got tweeted in isolation, which sparked the debate.
Who criticised this approach, and why?
Who disagreed with it? Pretty much anyone who knows anything about maps on the internet, among them people like Martin Stabe who I’d always consider a much better journalist than I am. It made me think twice about it, so I blogged to explain my reasoning. But there was also I think an element of me feeling a bit defensive, that the onslaught of “you shouldn’t do maps like that” was getting directed at the Ampp3d account at a time when it was being run by someone who hadn’t made the map, and they were copping all the flak from my design decision.
Would you say there is a tension within data journalism between accuracy and clarity?
Yes. I think it's obvious that I made the wrong decision on this map, and I doubt we’ll use it again. But it isn’t going to stop me thinking that part of the experiment with Ampp3d is to find out how you can tell interesting stories with complex data in a way that works on a tiny, tiny screen. I’ve got a different idea about how we’ll do it next time.
Look at Google News and every day you can see thousands of articles published by news organisations with no links to sources, incorrect use of data, poor visualisation, and things that simply don’t work on phones. We’re trying to buck the trend here, and we won’t get it right all the time. I appreciate that people pick up on when we’ve tried something that doesn’t work, or where we’ve made mistakes.
Click here to view the original article on Ampp3d, and check out Martin Belam's blog for further discussion on the debate about straight lines.
No comments:
Post a Comment