The Data Queen speaks exclusively to Claire Miller, senior data journalist at Trinity Mirror, on her experiences on producing quality journalism with data and her top tips on the best tools of the trade to use.
Data Queen: How important is data journalism in the newsroom?
Claire Miller: In some ways more and more so, there is much more data being gathered on all kinds of things (governments love to measure performance) and more and more of it is available and useable making data journalism skills increasingly useful for finding stories.
It doesn’t replace other kinds of journalism but it is something journalist needs to consider along with things like social media/community management skills and for finding new ways of working with your communities, all of this is part of your skill set.
Data Queen: What has been the most interesting data story you wrote and why?
Claire Miller: It is terrible to say they have all been interesting, that is partly because I cannot think of one in particular, but the reason I love doing data journalism is because I am always finding out interesting things, like where one in 10 pupils have been excluded from school more than once or that two-thirds of non-working parents in parts of the home counties say they are staying home to look after their kids, while their partners work compared to less than half in some places in the North.
That is one of the good things about data, you find out interesting things, which help when you are spending hours cleaning up messy datasets.
Data Queen: With some data stories journalists sometimes just tend to produce a lot of statistics and call it data journalism, what practical tips can you give to aspiring journalists to help them stand out from the crowed?
Claire Miller: Think like a journalist first – what’s the story?, what’s interesting? No one wants to read lots and lots of statistics (put it in later paragraphs, or an interactive so people can read it if they want and check what you’re saying makes sense), they want to know what is new that you’ve found out from the statistics. The data journalism supports the story rather than being the story. Data journalism is a tool to tell great stories, so tell great stories.
Data Queen: As data journalists, do you think we should be critical of other data stories, if so why?
Claire Miller: With data stories, as a general rule, if you can see what data they’re using, either because it’s publicly available or because they’ve published the background data, you can see how they are making use of the data to tell the story. It should be good practice to do that. From there I think it’s a good idea to be critical of stories that make unsupported leaps from the data or use very poor data as a source but I’m not sure every data journalism story needs to include high level statistical analysis to have value, knowing the underlying principles of things like significance and confidence intervals are useful, applying them to every story is probably not necessary.
DQ: What recent data story have you read, that didn’t work so well, in your opinion and why?
CM: Where things don’t work is where the data is under-used or misinterpreted, so stories that don’t adjust for population size – yes, there are more ASBOs issued to teenagers in London, there are also more teenagers overall – or don’t common sense check their data – near a police station is probably a hot spot for drug offences because people are searched at the police station when they’re arrested. Things like that are annoying because they should be things you check and they undermine your story (even if there is something there that is worth writing about).
DQ: What are some of your top tips for sourcing data?
CM: Check the obvious sources – The Office of National Statistics (ONS) publishes new datasets frequently and is the best way of keeping track of the latest ones. Parliamentary questions are sometimes useful, you can set up alerts for things you’re interested in through They Work for You, and put in FOI request.
If you want data, think of the best way to get it, either asking for database extracts or asking questions you can easily feed into a spreadsheet, it helps save time when it comes to compiling the data, so you’re not faced with thousands of pages of info.
But also keep an eye out for datasets and ideas. The Government and councils now publish a lot of open data and quite a bit gets missed, it is out there buried on data.gov.uk most likely, so spending time looking at those datasets and thinking about what you could do with them could be a useful exercise.
That and reading news is always a good idea, but I’m a journalist so I would say that, mostly for looking out for ideas for FOIs to put in or topics you could go and find datasets on. Good search skills and you’ll need them for the ONS website are probably an increasingly important tool for modern journalists.
DQ: What advice could you give to current and new data journalists on producing great data journalism and getting it right first time?
CM: Concentrate on getting the basics right:
- source data from places where at least some of the limitations will be known, all datasets have problems but places like the ONS at least tend to have an idea of what the problems with theirs are.
- if you spend time cleaning data, and you will, make sure you’re not creating more problems that you’re solving, such as deleting rows or getting things out of sync with sorts.
- You only need some basic formulas,percentage change, rates per 10,000 etc, percentage of, to get started.
- visualisations are great for telling a story – but make sure they actually tell the story you want, simpler is better so it’s not confusing.