Michael Goodier (Interactive Journalism, 2017) is a data journalist with Reach’s award-winning data unit. We caught up with him for a quick q&a regarding all things data.
Why did you choose to specialise in data journalism, and how did you get into it?
While my role primarily involves working with data, I would still consider myself a news reporter – just one who mainly uses data to find or reinforce stories. I’d done the odd bit of reporting with data at a local level before I did an MA course, but I mainly started to properly think of it as a useful avenue to find stories when on the course. The draw of data is that there is an absolute wealth of information out there, which can be used to find stories as well as reinforce stories that you might find through more ‘traditional’ means. I also enjoy the more investigative side of ‘data journalism’ – a well-aimed FOI request or a web scrape is often the start of a great story. After my journalism course I managed to land an internship with the Reach Data Unit (through the Google News Fellowship scheme – this year’s applications are now open). Luckily a place opened up, and they liked my ideas enough to keep me on afterwards!
How do you think data stories on a local level compare to those on a wider level?
That’s an interesting question. I would say that the main benefit when it comes to national data is the greater availability of ‘open data’ like government statistics compared to locally (although you would be surprised at how much is out there that breaks down to a local authority level or below). On the other hand, local and more granular data is far better as you can do a lot more with it. Ultimately a good local story will often also make a good national story. If your data is local enough, it can enable follow-up on-the-ground reporting, like this story which started as granular police figures on antisocial behaviour but was greatly enhanced by a reporter going to the affected areas and interviewing people. Looking just at a national figure on antisocial behaviour, you wouldn’t have had that detail.
Another example could be that you might find underneath a national rise in a figure, certain parts of the country are seeing a fall – which obviously works as a story for that local area, but also potentially as a national story focusing on why that divide might exist. You might also find quirks in the data that you wouldn’t when just looking at an overall figure – for example this story about one guy who was raising hundreds of issues with his local council on FixMyStreet. I believe that readers also connect more with stories that are local to them – and interactives that allow readers to enter their postcode offer a personalised way in to a story that might otherwise be a bit dry and stats-heavy (shout out to our development duo Carlos and Cullen who do the database and javascript side of these).
As a collection of regional titles, how exactly does the data unit at Reach PLC work?
So the bread and butter of what we do is our local data bulletins. These are local versions of stories that get sent out to all of our titles (where there is a story) as well as a national version which gets sent to the Mirror. Usually this is either based on open data, data we have scraped from the web or APIs, or a Freedom of Information request. Sometimes if something is in the national news cycle that day we will also do a quick turnaround to get out local angles for our regional titles. We will write a story based on the figures and then get a comment or reaction that explains or adds something to the figures and also works nationally. It is then up to the local titles to get local comment and case studies – as we write for 40 titles (and that figure is constantly growing) it would be very difficult to get local comment for everywhere on top of everything else.
The second main thing we do is write national stories that are sent out to all of our print titles daily. The focus of these is normally an eye-catching graphic (usually made on illustrator by our designer Marianna Longo). I love these as it gives us that opportunity to cover the national and explainer style stories that data might just not be available for on a local level. We also send out graphics on basic stats for example crime, house prices and homelessness to our weekly titles that are generated using R and photoshop script.
What has been the most complex data story you’ve worked on and why?
Often the simpler stories are the ones that resonate the most with readers! You don’t necessarily have to use complex tools or techniques to come up with a good story – and complicated certainly doesn’t mean good when it comes to the final piece. That said, in terms of difficulty, some of the FOI-based stories I’ve done have been frustrating – for example a piece I did on Community Protection Notices last year took ages to compile as I had to FOI every council, and not all were very forthcoming.
In terms of programming, maybe the most difficult at the time was a big Companies House story I did. It involved scraping the net worth of every company, and using an API to guess the gender of boardroom members from their name to come up with an “average” boardroom for each area – I was still getting to grips with R at the time so it definitely pushed my skills. Maybe this is a bit of a cop-out – but different stories are challenging in different ways, and the thing about being a data journalist is that your specialism is a method of journalism, not a subject area, so you are constantly learning about different things, be it accounts data, justice statistics, or the ins and outs of how different figures are recorded on government databases.
What tools do you use the most in your data reporting? (e.g. what training did you have at Reach PLC, did you self-teach any coding languages?)
I love R and the tidyverse – I was taught the basics by big Basile Simon during my masters and I now use it pretty much every day: to web scrape, clean data, interrogate APIs, do geospatial stuff, make maps and charts, and sometimes even to write the skeleton drafts of stories with. I knew a bit going into my role, but mostly it has been a journey of self-teaching on the job through problem-solving (stackoverflow is obviously always a great help), and I now feel pretty comfortable doing most things with the language. Of course the tool I use the most is probably a spreadsheet programme, and a word processor.
What advice would you give to aspiring data journalists?
The usual – and best – advice on doing the job itself is that the data is never usually enough on its own, it’s often what you do with it and what else it leads on to that makes a story. Pretty much everything that applies to a ‘normal’ reporter also applies to a data journalist. Ideas are the main currency in data journalism – and that is also true for journalism more generally. You can approach coming up with an idea from several directions: either starting from an issue you are aware of then finding data on it, a tipoff that something might be happening and sending off some FOIs to build your own dataset, or from the data itself – spotting an interesting quirk in the numbers, combining multiple datasets to come up with a unique angle, or, for example, thinking about which angles would work and what you would gain if you scraped a particular website.
In terms of getting pieces commissioned or performing well in job interviews ideas are also key – I have been turned down for a job before because I was going through a bit of a mind-block and didn’t bring enough ideas to the table. My career advice – not that I am an expert – would be don’t be an ‘aspiring’ journalist. Pitch your work and do the journalism anyway, and that way you have something to show for yourself. Other tips – be very specific when writing FOIs, and always check disclosure logs and WhatDoTheyKnow (I have had countless stories and ideas from FOIs that have just been sitting there). Also pick up the phone – it’s quicker and you will often get better quotes!