“You couldn’t work as a journalist, if you were not able to do an interview. The same applies to data journalism in the age of digitalization” – says Nils Mulvad, a world renowned data journalist, editor at Kaas & Mulvad and associate professor at The Danish School of Media and Journalism during the Data Harvest 2014 conference.
Paulina Pacuła from European Journalism Observatory has conducted this interview with Nils Mulvad. We bring it with her permission. The interview was published in Polish here. An edited version in English can be written here.
For many journalists data journalism is basically about making data look nice – visualising it, creating interactive charts etc. But that’s not all, is it?
Definitely, it’s not all. Data journalism is an early alert tool. By analysing data you can sometimes recognize problems, before they actually cause a lot of harm. Or before you are able to notice them using other journalistic tools.
To give a quite recent example: the largest financial pyramid in the history of US, created by Bernard Madoff, was exposed in 1999, years before the big scandal happened. The man behind it was financial analyst Harry Markopolos. He informed the Security Exchange Commission, that he believed it was legally and mathematically impossible to achieve the gains Madoff claimed to deliver. Markopolos came to these conclusions through very comprehensive data analysis. That was the job journalists should have been doing.
This is very arduous work to do.
Using data journalism tools like spread sheets, scraping and data visualisation makes it much easier. First off all data journalism provides tools for continuous research. Scraping allows you to monitor how different institutions work on a daily basis, because it allows you to monitor new stuff showing up on different webpages.
On the other hand data journalism tools make it possible to work with huge amounts of data, which is very important. Especially today, when public institutions produce tons of data accessible to journalists through freedom of information laws. It would be difficult to deal with farm subsidies or EU tender data only using your brain and calculating skills. Today we have great software, that allows us to analyse thousands of records.
Data journalism became this big trend only a couple of years ago. But in fact this field is much older.
Yes. Journalism was always about data analysis. The Wall Street Journal was born in 1889 from a daily Dow Jones afternoon letter about the stock exchange. In sports reporting, data has always been an essential aspect of this. John Snow’s map of cholera outbreaks from nineteenth century London changed how we saw a disease – and gave data journalists a model of how to work today.
But the data journalism we know today, started with something called Computer-Assisted Reporting. It was back in the 50s, when reporters started to analyse data using software tools for social and scientific analyses. One of the first examples of computer-assisted reporting was in 1952, when CBS television used a UNIVAC I computer to analyze returns from the U.S. presidential election.
At the beginning there was a lot of excitement about the tools. Journalists would report in their stories, what kind of computers they used, how many records they had analysed etc. It’s like today you would say: after doing five interviews on iPhone 5, I came up with this and that conclusion (laughs). Back then, the problem was that very often CAR specialists would only focus on getting the data, instead of getting stories out of it. So basically it was about showing the readers: ‘look, we can analyse so much data’, instead of really using it as a tool.
Databases became central to journalists’ work by the 1980s, because of the ongoing process of digitalization. There have been also some new jobs emerging in newsrooms for people mainly focused on presenting the data, visualizing it. Today data journalism is one of the most important trends in the media world.
Professional data journalism requires lots of skills, which were not associated with journalism before. For example nobody required journalists to understand programming and coding.
Yes, that’s true, but this is not exactly the case today either. Of course, in order to do data journalism, you need to combine three types of skills – journalistic, programming and web designing. Sometimes you will have them combined in one person and sometimes you need the whole team. But it is necessary, that people understand other fields of work, because only then can they communicate and work effectively. If you know nothing about coding, it will be difficult for you to work together with a coder, because you will not know, what he or she needs to do the job. So you need to have knowledge in all three areas and be at least very good in one of them.
I would say that there is another type of skill required: data analysis which comes from social sciences.
Yes, of course. You need to understand basics of statistics and other social science methods. One needs to be very careful with data and use it in a proper way and not come out with conclusions which cannot be supported. That’s why data journalism is also called evidence based reporting.
There is this quality associated with data: facts. Data mean facts, something objective. But isn’t true that data driven journalism is still opinion-based journalism? There is always a certain level of subjectivity in how the story is shaped and presented, what conclusions are derived from the data.
The more the story is documented and researched, the closer to reality it gets, so data journalism is more about facts than opinions.
Very often data is only a first step in the process of creating the story. But it is very important, that the journalist looks into the data himself, instead of relying on other researchers. This is something journalists do too often: they interview people who know the data. But those people may have their hidden agenda, they may have the conclusions a bit coloured, in order to justify the meaning of their work. Journalists should be able to find the most important stories in the data and then interview sources on their findings.
By the way – I don’t like the word ‘data driven journalism’. It sounds like the data is the most important part of the job. We see a tendency to underestimate journalism. This is what I was talking about at the beginning. These machines are not the most important part of the process . These are just tools. But it’s still the journalists and their ability to think critically, which are the most important part of the job.
If you look at data journalism from an academic point of view, how does it change journalism? Is it a big shift?
I think data journalism itself is not a big shift. There are a lot of new tools emerging all the time and they only change the methods of work, not the work itself. If most of the information today is in digital form, you, as a journalist, should be able to gather it and analyse it. It’s the same as when you gather information by doing interviews. You couldn’t work as a journalist, if you weren’t able to do an interview. Data journalism is the same – it’s just interviewing data sets instead of people.
But yes, we are in a period of change and it has a lot to do with things going digital. The business models of many media, especially print ones, aren’t working well. Many media companies need to reinvent their business models, but they are very reluctant to do that. It seems like they are waiting for the internet to go away, but that’s not going to happen.
What we see is a shift from institutional media to more socially based media. Journalists are getting much more independent thanks to cross border cooperation opportunities, crowd funding and the emergence of new communicative platforms. Those who have difficult times in media organizations, which are trying to act the old way, can either walk away, or try to introduce changes, instead of knocking their heads on the walls. If you are a good journalist, people will be most interested in following your work, instead of the media institution you are representing. Sometimes it’s easier to attract readers to a whole new website, instead of to the old media.
And at this point you see another important part of what data journalism is about. It’s also about new ways of spreading your stories.
Exactly. The “one to many” model of communication is coming to an end. Now it’s more like “many to many”. Journalism doesn’t finish the moment you have prepared the story and published it; the next step is to get it out to people. The stories should be well targeted in order to create an impact. If you are not capable of spreading your story, it means it’s only written for the archive.
The skills of building a community, identifying key influencers, showing how you write the story, forcing media to disseminate it – this is also very important in the new media landscape. Data journalism is a mixture of all these things. Of using all these new tools – also social media – in a professional, journalistic way.
There has been interesting research done on the use of social media by journalists in Poland. What is peculiar, is that they don’t really treat social media as a possible source of information. They use it mostly for “inspiration” instead of proper research – finding people, facts and numbers. Is there something we are missing about these new platforms of communication?
Yes, definitely. Social media can be used as a tracking tool for social relations, which is very useful in investigative journalism. There are some examples of investigative stories based on analyses of social circles on Facebook or LinkedIn. One girl from Slovakia actually managed to track possible corruption cases in health service procurements in Slovakia by analysing all the contracts and personal networks between hospitals’ boards and the management from the companies taking part in the procurements (link). People leave a lot of data and information about themselves in social media, so it’s good to know how to follow this.
But research is not all. Social media are very powerful tools for spreading your stories, interacting with readers, creating impact. And it is quite funny how many media do it. Journalists don’t get involved in discussions; don’t interfere with the audience. Once they post the story, that’s it. This is totally wrong.
You should think about social media as like a party. When you come to the party and people are talking to you, but you are not responding, they think you are rather impolite. If the only thing you do at that party is sometimes stand up in the middle of the room and shout, what you got to say, and then, when some people are approaching you to discuss this, you say ‘oh no, I’m not going to get into a discussion, they will probably think you are crazy or arrogant.
There is a lot of this kind of craziness or arrogance in traditional media. But this model is coming to an end. If you don’t know how to treat your reader, you’re going to lose him.
Who should we look to for best practices?
Well, I would say the Guardian is doing a very good job, New York Times, Los Angeles Times. Norwegian press is very good in adapting to changes. Danish Broadcast Corporation has also set up a database analysis team and they are doing some good stuff. But these are just some examples, which come to my head right now.
But looking for best practices in data journalism, you should also look at many cross border, independent networks, for example ICIJ – International Consortium of Investigative Journalism. The best example is their tax haven story. They’ve got access to some data about people having accounts and companies in tax heavens. They teamed up with different media all over the world and gave them access to the data for that country. The analysis took half a year. They have decided to let the stories out in three waves to create a bigger impact. They exposed the biggest CEOs from German banks, the wealthiest people in China, UK etc.
It was a very big thing last year. They’ve got a lot of awards for that. In effect the EU started to work over anti-money laundering rules. It could not have been achieved without using the tools of data journalism. What was also important in this project, was the international cooperation between journalists. There was a deep expertise needed from each country involved, so it had to be done by international group. But today, with such tools as social media, free software everything is possible. Data journalism is a sort of journalistic punk of our times.