Mirko Lorenz defines himself as a journalist, information architect and educator. He teaches journalists the skills of data driven journalism. A project he currently works on is Datawrapper.de, a tool for easy creation and publishing of data visualisations, increasingly used by newsrooms. He wants to create systems and web sites which are based on data and journalistic content and enable a better experience for users.
You are experienced as data journalism trainer, including in Bosnia and Herzegovina. What do you focus on in your trainings, i.e. what do you teach journalists?
The focus is on building an understanding that data is about story. That in many instances understanding and reporting based on numbers gets you a stronger case. What I found, though, is that this is a new world for many journalists. They never did research in the statistical offerings, they have a hard time with tools like Excel, and they are undertrained how to visualize data effectively and correctly. And at the very end: they do not see yet that trustable, data-driven reporting could be a future - for writers/reporters, as well as for media organizations.
What do you find to be the biggest challenge for journalists as they learn data journalism skills and tools?
If you walk into a forest the first time, everything is confusing. If you would walk through it every day, it would be easy to spot differences. This is a fitting picture for what is data-driven journalism now: a new world, which is a bit of a surprise. However, the practice can be learned - people get it once you show it. My belief is that we need to be patient. At best, we can create a process where - by experience and by success - data-driven reporting becomes an accepted process in the newsroom.
Tools, by the way, are interchangeable in many ways. If you don't have a plan, every tool can actually be as harmful as it can be a solution. Master Excel, learn the rules for correct visualizations and avoid the pitfalls of false correlations (if this, then that - make sure that there is a connection really). Be aware that the tools themselves do not help that much to come up with great data insights. Instead be an inquisitive mind. Ask questions. Ask questions others want to have answered. Then start digging into the problem, mainly look at available data, not as in the past just asking a few people. Write a first story about the (potentially) missing data or the bad data quality. Often that is the case.
I think a good data journalist is like a good photographer - more important than the camera is the "eye" for the right shot, the right moment and, of course, a certain mastery of technicalities. The rest comes with just doing it - a lot.
How would you define data journalism as it is now and how do you see it developing in the future?
Quoting a fun fact from Nate Silver's book: well into the 90s the CIA was assuming that the GDP of the Soviet Union was about half of the GDP of the US. Even at that time that was neither true nor correct. But a lot of political actions were based on that false data. In 2008, the subprime crisis in the US was based to a large extent on false models used by the rating agencies. I could go on - there are many more examples.
People, including you and me, make a lot of decisions based on assumptions about the world. Sometimes we listen to friends or family. But looking into the future it is clearly possible to help people make better decisions, to keep them from getting ripped off, to avoid big confusion in big debates.
Data journalism or data-driven journalism (which I prefer) is essentially good storytelling. It is saying that this or that has changed and how it might affect you. Readers have a good sense for what is just a filler of information and what is more valuable, deeper stuff. They know it, they sense it. From a newsroom perspective it is about ´knowing better´ and being able to proof it. Say, your local government is pushing for a new pension plan. Can your newsroom really judge it? Data-driven journalism, from my point of view, is just good journalism for the 21st century.
Are there data journalism skills that every journalist should have nowadays and why is that the case?
Ideally we would have a number of journalism-coders in every newsroom. But I do not believe that really every journalist should become a hacker. Instead what is really important is a sense and experience for the many fields and steps needed to analyze data: we need excellent researchers, great visualizers, forceful writers. The main skill I expect from a journalist is that she or he has not given up. All the rest is trainable, even in weeks or months.
How should a self-taught journalist begin to learn data journalism? What is your advice for resources, tools and steps in the process? Where to start from?
Try to assess what you really like to do. If you have a knack for programming, start using resources like Code Academy or, even better, team up with others who already do that. Try to create a first project that is based on data and has a "SLAT" (BBC jargon for "shit, look at that"). Then do it again.
If you are feeling that programming is not for you, become knowledgeable about researching data, about using tools like Datawrapper or others for data-driven journalism without coding.
Start small, do it often and focus on shocking everyone with surprising facts and visualizations. If that works, no one will need to train you. Instead, and a lot of journalists are like that, you will dig yourself through what is there to learn.
What is your experience with newsrooms ushering in data journalism skills? What would be the approach for reluctant newsroom management to motivate its staff to turn towards data journalism skills?
Most newsrooms are broken. They have a failing model, but resort to simply reporting faster and more shallow. Many, including some brands that would never have considered that in the past, use elements of the yellow press. Quoting from a book by Stephen Few ("Show me the numbers"): The problem are the bosses. If a newsroom would publish text with typos, there would be consequences. If a newsroom publishes wrong data in wrong visualizations, nobody cares.
It might sound cynical, but I do believe that the crisis must become even worse before that will really change.
You define yourself on your website as 'information architect'. What is an 'information architect'?
I am part journalist, part developer, though no coder. I start my projects with a blank page and then develop websites and topics. So, in my work I have to consider technology, user experience and content - and that is what IA (for "Information Architecture") is basically about. I feel that it describes what I do better than saying that I am just a journalist.
Datawrapper.de is one of the projects you are currently working on. Tell us about it. Also, tell us a bit about other projects that have been occupying your time recently.
Datawrapper aims to simplify the process of creating and publishing simple, correct charts. It is an entry level tool for data-driven journalism. We want to get journalists to work with numbers. And it works quite well – we have more than 2 million visits from embeds all over the world. Right now the Guardian, Le Monde, the Washington Post, Frankfurter Allgemeine Zeitung are using the tool, on a daily basis or from time to time. As for my time: I work in two shifts per day. One is the normal job, another one is at night and on weekends. But no worries, I like what I do and I do stuff other than working. And the progress in #ddj (Twitter) is a big motivation.
Although data-driven journalism is my focus, I work in a range of other, somehow related projects. One big project, now in its third year, is VISION Cloud, an EU-funded research project for future cloud technologies (http://www.visioncloud.eu). The big, big plan for 2013 is to work on what would be needed to build a data desk for a newsroom - with Datawrapper, but with other tools, too.