[This is a special guest post by Claire Miller, about the data journalism being done for the Trinity Mirror regionals. Claire was among the shortlisted for Best Individual Portfolio in the 2015 Data Journalism Awards. Disclaimer: Infogr.am, the company behind Visualoop, was a sponsor of the 2015 GEN Summit]
I work for Trinity Mirror regionals as part of their data unit. I’m part of a small team of five journalists, a designer and a coder. Our job is to take data we find, from government statistical releases, freedom of information requests, open data etc. and turn it into potential stories or interactives the regional papers and the Mirror can use to inform, and sometimes entertain readers.
The data unit grew partly out of work I’d been doing in Cardiff using data. I’ve always enjoyed maths and had a tendency to pick up statistics related story right from when I started as a journalist. When I moved to Cardiff, I moved from patch based reporting, where the starting point for finding stories is at least somewhat clear, to general reporting in a place where I had no contacts. So in order to have something to put on the news list each day, I started digging through the ONS data releases.
This was around the time the Guardian Datablog was taking off and knowledge and tools for doing data journalism were becoming better known and more accessible, so I decided to embrace it and see if it was possible to apply techniques and tools such as Tableau and Google Fusion Tables to regional news (turns out the answer there is yes).
Starting out we tended to use free visualisation tools, like Fusion Table maps, Tableau, Datawrapper and Infogr.am. Now we have the advantage of having a coder and a designer, which gives us a lot more flexibility to create things.
The pages for each school in the Real Schools Guide (http://www.gazettelive.co.uk/all-about/primary-school-ratings) use the Google Charts API, using a Google Fusion Table as a database, to generate the individual charts for each school.
The graphs are simple, bar charts and line graphs to show the background information the guide is based on – for example, the proportion of pupils at a school getting five or more GCSEs at A* to C and how this compares to the local and national averages. Sometimes, just a simple graph will communicate the figures, and the chart API is a useful tool for generating lots quickly.
There are issues with trying to create visualisations simply and for free. There seem to be less free tools available now, as several have closed or now have any limited free use, although we may just be in a lull before some new ones are created.
Another big shift has been in terms of how mobile focused we are now, at some parts of the day the majority of traffic to the websites is mobile, so anything we build has to look and work great on mobile. When I started it, it was nice if it did but it wasn’t the major focus.
Quite a few of the free tools, Tableau for example, don’t always work that well on mobile screens, anything with maps risks readers getting stuck scrolling the map not the article. It makes it more difficult to create things you know readers will be able to use.
Being part of a team with designer and a coder means we can build our own visualisations that work well on both mobile and desktop.
In terms of creating visualisations, beside the ‘it has to work on mobile rule’, we also tend to aim for things that are clean and simple to use and understand, and a lot of the time we’re trying to build things quite quickly.
The interactive for the Welsh Index of Multiple Deprivation is a good example of taking something complicated and making it useful for our readers.
The index is published every three years by the Welsh Government. It is a comprehensive analysis of the relative levels of deprivation in different small areas of Wales. As with most statistical releases, the data and reports were published at 9.30am, meaning a need to work quickly to get the data analysis, as it was likely to be one of the main breaking new stories of the day.
The interactive news app, built by Carlos Nóvoa, is the focus of this, allowing people to find out where their area ranks in the index, giving them an idea of how deprived where they live is on the eight areas focused on by the Welsh Government, such as employment, health and education.
We picked the format, with scales showing where the area falls relative to deprivation in other areas over a Google Choropleth map format, which is the usual way of showing information for different geographic areas, as it meant we could clearly display much more information than we could in a map, such as the breakdown of the area’s rank by different measures, which would not have fitted in the small info bubbles needed for mobile.
The design by Dmitri Thompson is clear, easy to understand and engaging, and helps to emphasise the relative nature of the index, which ranks different areas according to how their deprivation level compares to all other areas.
The index uses Lower Super Output Areas (LSOAs) for its lowest level of geography. LSOAs are a common unit of statistical geography but they often aren’t area boundaries that people living in them would recognise, unlike council areas or wards. The challenge with this data was to make it personal in a way that people could understand, which is what the postcode search does, it brings the data back to the level of where someone lives, works or goes to school, helping them understand what that area is like, without having to have a technical knowledge of the statistical geographies.
We try to reuse interactives where possible, re-purposing one to show different information. So the search tool to help people find out how good their local GP is was turned into a search tool to allow people to view how well their local school did at GCSEs and A Levels.
This re-purposing allows us to create things quicker, by taking out some of the early designing and building stages.
*Claire Miller is a Senior Data Journalist working for the Trinity Mirror Regionals. She’s also the author of Getting Started with Data Journalism – Writing data stories in any size newsroom. You can follow her updates on Twitter (@clairemilleruk)