"People have all sorts of illusions about investigative journalism which I guess I wanted to tackle"

(Photo by Dawood Hilmandi. Creative Commons licence Attribution-NoDerivs.)

How do you practise data journalism in your day-to-day work?

Data journalism is just one among many possible options, but it tends to be a tool I use a lot for a number of reasons. Firstly, because it's something that I like doing and it tends to get you stories that others wouldn't get, because not many people have those skills. It's also a way to get you stories about important issues, whilst at the same time enabling you to tell those stories in ways that are compelling. A good story should have two ingredients: it should tell me why I should care and why this matters, it should have a human element and a side that digs deeper and tries to explain things. That's why I turn to data quite a lot, and why I think it's a skill worth learning.

Data journalism is in your view not an end in itself?

Exactly: the story at the end doens't even have to have any numbers in it. The data can lead you to the story and you can spend five minutes on the numbers or even weeks, but a lot of the skill in data journalism is in knowing when to turn away from the data and move on, because it's very easy to get distracted and to get bogged down. You need to be very focused, because often there are so many stories in the data.

Are there any recurrent mistakes you see in online data-related publications? If so, how could they be avoided?

There's often too much emphasis on visualisation. One particular mistake I often see is the creation of maps because the data is geographical - you should use a map if it tells a story, or allows a user to find one. But if it's not doing that - showing a clear pattern of distribution, for example, or allowing us to find our local takeaway's hygiene rating - then use a chart instead. Or just write a story with text.
Related to that is too much focus on numbers. Data can be the start of a story, but numbers often aren't the best way of telling it. Find a person that those numbers affect, or who is responsible for them. The numbers should be about people, and the story should too.

Data has a bad reputation in some newsrooms: it seems time-consuming and resource-intensive. Rightly so?

No. It should save time. I've ranted about this on my blog.

You've worked a great deal on crowdsourcing projects where the public in no small part contributes to an investigation. What's the thinking behind that?

The reason that I started Help Me Investigate was that I wrote a book chapter about investigative journalism online and in the process of researching that chapter, I found a number of examples of crowdsouring which, it seemed to me, adressed some of the challenges that investigative journalism often faces, like subject choice and impact. And crowdsourcing seemed to be doing very well on those fronts. Crowdsourced investigations seemed to have a massive impact, compared to if you had done it in traditional ways.

Likewise, you had subjects being investigated that a journalism institution wouldn't have investigated for commercial reasons, for cultural reasons, and so on. Investigative journalism goes through fashions, and certain subjects go out of style, if you will, but that doesn't mean that these issues have disappeared from society, it just means that the news agenda has changed. Crowdsourcing was therefore quite an interesting way of broadening the news agenda and doing different types of investigations. So basically Help Me Investigate was about a) doing different types of investigative journalism and b) having a different sort of impact, but most of all the aim was to make it easier, or cheaper if you like, to do investigative journalism.

Also, one of the things that struck me about investigative journalism, is that it's almost like a magic circle: first of all you have to be an "investigative journalist", which draws this line between some journalists and others, which I think is horrible. There are investigative journalists, and they should be celebrated but the term can be exclusive and it makes some people think they can't do investigations. They think it's some magic art and it isn't. It's largely about persistence and curiosity and having questions and looking for answers. People have all sorts of illusions about investigative journalism which I guess I wanted to tackle.

Finally, a large part of its success is the social dynamic it creates: if you are a journalist, or a student journalist, or a non-journalist, on your own wanting to ask a question, then you probably won't pursue it, but if you can find one other person that wants to ask that question, that social dynamic makes you much more likely to do something about it, because you know someone else is relying on you. So if nothing else, it's about putting people with different skills, backgrounds and experiences in touch with each other and when it works well you get something that's really special and you get a story that no news organisation would be able to do alone.

How do you motivate voluntary contributors and ensure the quality of their input?

Money isn't the only motivation.  First of all, it has to be something the contributors are interested in themselves and want to do. It's not me saying "I want to investigate this, who wants to help me?", it's rather someone else coming forward with an idea for an investigation and me replying "How can I help you?" There has to be a shared agenda to get more people involved. And they have to get something out of it other than money, whether it's gaining skills, or contacts, or something else. But most of the time just gaining an answer to a question is sufficient. People contributing often don't want to write a story at all, just get the answer they're looking for.

As to the quality of the information: the type of crowdsourcing we do, is largely done in small groups, that way it's easier to establish quality and trust. And we do the same kind of checks as other news organisations. The only difference is that we're not all in the same room at the same time.

Over the years, you've done quite a bit of experimenting with online journalism. Why is that important to you?

It started when I began teaching Online Journalism, because if you're going to be teaching that, you should be trying out all sorts of things. Experimenting is also important because no one really knows what are the rules of the online game, what are the best practices and so on and that's going to continue to be the case for a good ten or twenty years. We're still inventing the rules of what works and every five years it changes again because mobile comes in, or tablets come in and the next thing will be other devices. But after a while you can start connecting the results and insights of one thing with another and start gaining an understanding into what actually works.

What is a good way to see changes coming in online media and stay ahead of them?

Follow people and news sources blogging about developments in the field: there are good general news sources now like PaidContent and Journalism.co.uk, but also look for the more specialist ones like Search Engine Watch (SEO), Feverbee (community management), multimedia journalists, and so on.

Watch out for new tools, have a play, and move on. When you need them you'll think of them - don't get too hung up on mastering everything, again - I think it's more about being curious and knowing what's out there.

Do you have any tips for people getting started in online and/or data journalism? What skills should people acquire, tools and programs to master?

I always find that people learn better by finding a problem they want to solve - a question they want to answer - than to seek to master a skill alone. There are so many skills you could learn in data journalism that it would be very easy to lose motivation before you've even begun. So firstly, don't let that overwhelm you: everyone feels the same way (including me). Start with a problem, and stay focused on that. If it becomes too hard to answer, choose another one. That's the process I use as the basis of Scraping for Journalists, and the ebooks I'm working on now.

The skills you'll probably need to solve a problem are likely to start with: using spreadsheets to make calculations and create pivot tables (aggregate overviews of the data); and using data visualisation tools (Datawrapper is a good one to start with; or Batchgeo for maps). But if you find a different problem that requires a different skill - say, FOI, or scraping - then go for it.

What in your view are subjects that merit investigation, what are the stories worth pursuing these days?

I don't think anyone can dictate what others should investigate. We should be wanting to get as many people investigating what they can. So I guess the stories worth pursuing are the ones that you can pursue, not some abstract image of being the 'investigative journalist', which excludes more people than it inspires. The stories that you care about, and the stories that no one else is telling. Whether that's pet food or local schools or international war crimes, I don't think anyone should try to rank the results as more or less worthy.

What's next for you?

I just finished Scraping for Journalists, which was basically programming for journalists. I now have a good 4 other books that I want to write. Three that are similar to scraping for journalists in that they are about particular aspects of the process. First: cleaning data, which is a whole skill in itself. Then spreadsheets for journalists, and data visualisation. But before that I want to do a short book which is based on a one-hour data training session that I do. It's just about the process, and kind of how to get started: dealing with an easy dataset that no one else is looking at, finding stories quickly and then just moving on to the phone calls. It's about the idea that data journalism isn't only about spending hours on a spreadsheet. You might spend five minutes on a spreadsheet and the real work is in the phone calls and chasing the reasons behind those numbers.

By Arthur Debruyne