There seem to be a few common trends among successfully integrated teams. Firstly, an acknowledgment at all levels within the newsroom that technical literacy and technical skills can open the door to new forms of storytelling. This could manifest itself in ways such as the data team being invited to regular editorial meetings, just as the “traditional” journalists are — or even in the way that they are referred to, with “journalist” in their job title.

Secondly, newsroom managers need to recognize that the data journalists aren’t there to solve everyday technical issues, even though their technical know-how might mean that they are able to. Nowadays, a newsroom has all sorts of digital demands that add an extra layer of technical complexity to a regular organization, such as publishing online, or having a suitable content management system (CMS). In addition to that, there might be other more regular technical issues that arise — from digital security practices to websites going down, or bugs appearing.

In most cases, though, addressing these issues simply doesn’t lie within the mandate of a data journalist — and though it might be quicker if they’re sitting at the desk just around the corner, this can seriously undermine their journalistic responsibilities. In much the same way that any random native English speaker wouldn’t be asked to copyedit a piece before publishing purely on merit of their language skills, an individual who has a certain technical skill set is not necessarily the go-to person for all of the problems that involve digital technologies.

Respecting the diversity of skills that lie within digital technologies is crucial for many reasons. For the data journalist or team in question, it gives them the time and space they need to do the job they signed up to do — telling stories with data. For their peers, it can send a clear signal that the data journalists are not simply the IT department or the system administrators with a different name, but that they are peers in the field of journalism. Knowing that they are there to go to with ideas or questions can (and should) bring up all sorts of collaborations that can make a story that much stronger.

Thirdly, news organizations need to prioritize communication, and understand that it might not be that easy. For those who aren’t sure what benefits data-driven journalism could bring to a story they are working on, flagging it as early as possible with the data or graphics team leaves space for new approaches to the story. For those with technical skills, sharing knowledge can be a good way of flagging to colleagues areas that they are interested in, as well as bringing the broader benefit of boosting data literacy across the newsroom.

Avoiding jargon in communication can be important too, and providing spaces where questions are welcomed rather than seen as a sign of a lack of knowledge. One example of this can be seen in the regular “learning lunches” that Noah Veltman, a developer placed within the BBC for a year as part of the Knight-Mozilla OpenNews fellowship, held on a regular basis. His colleagues were invited to drop in and learn about technical topics they might have heard referred to, and the sessions were well-documented for future reference.

To a degree, this means that those who really do understand the benefits of a data-driven approach are left with the task of advocating internally within their own newsrooms to make changes to age-old processes, and incorporate new forms of storytelling into their work. While perhaps the onus shouldn’t necessarily be on them to change these practices, pragmatically speaking, they are most likely to be the best placed to do this job, and the outcomes will, ultimately, benefit their positions as well as the newsroom culture as a whole.

You could argue that all the new tools for visualizations — Google Fusion, Tableau, CartoDB, Infogr.am, and others — have helped the lonely journalist to continue working on his/her own, but in one way I think that those tools have increased the collaborations between the data journalist and developers, graphical artists and designers. The tools helped journalists to realize that a story does not necessarily have to be a headline, picture and text, but you could actually tell a story in a lot of different ways. And once they realized that, they also realized that yes, the easy tools were good for those first stories, but very soon they wanted more: more and better interactive graphics, and they realized that there were people out there who actually knew how to tell stories in other ways.

An example is the project with property prices that the Guardian did during the summer of 2015. It started with a database. And that is, for me, an unusual way of starting a story. Usually you have a story idea and then you set out to find data that will help you with that idea. In this case, we had the data, and wanted to find the stories. The data is a public dataset, found here: https://www.gov.uk/government/statistical-data-sets/price-paid-data- downloads. It contains of 19 million records of property sales in England and Wales from 1995 onwards. The dataset had been used before, mostly by commercial sites for finding properties for sale or for rent. They had chosen to display the latest sales for each area, so that was not something we wanted to do. Instead we realized the potential in having all the data from 1995 — we could really tell a story about how house prices have developed over the years. You could argue that there is nothing new there: prices have gone up, we know that.

My job at the Guardian is to collaborate. In the work description it even says that I and my team should be “aggressively collaborative” with the news desk, graphics, the interactive team and so on. So, in this case, I lifted my head and started talking to the interactive team right away. But it wasn’t until I also talked to one of the reporters that I realized what the story was. He had an idea of writing about a family in Islington, London, where the parents had bought an apartment when they were young and just starting a family. And now when their children wanted to do the same thing, they couldn’t afford a single apartment in London. So, we joked that they would have to move to the northern parts of England, because we assumed that flats would be cheaper there.

I then got the idea to compare the median price with the median income over the years and asked the HMRC for data by region from 1995 and forward. So, for the different regions in England and Wales I could do a comparison that showed that in 1995 the median price for a property was between 3 and 4 times the median income for that region. So, London was most expensive then too, with 4.4 times and the North West was the least expensive with 3.2 times. But, when I looked at the latest data available, I got a surprise. London had of course risen the most, to 12 times the median income, but even in the North West you had to pay 6 times the median income for the median house. So, if there is an old rule that says that banks will lend you 3 to 4 times your income, it’s not possible to buy a property anywhere. And with that information, the story got a new angle – and we could start sketching on an interactive visualization.

So, I prepared the data and put it into a PostgreSQL database and our interactive team started to work. But the great thing is that both them, and me, can lift our heads at any time saying, “Could we add a field with postcode district?” or, “I’ve managed to find about 24,000 of the 30,000 missing postcodes, you can now re-run your script.”

To be able to work together has improved the project immensely, and we can see how the story angle changes because of the interactive, as well as the other way around. You can see the final outcome on the Guardian’s website here.

[One] reality for the every-student-must-code enthusiasts to consider is that the vast majority of journalism students simply aren’t interested in being high-level data journalists. I am blessed with really smart students, many of whom are in our joint Bachelors/Masters degree program. But it has been my experience in teaching hundreds of these students over the years that maybe one or two in each class of 20 might go on to do data journalism at a level beyond simple spreadsheet work.

I’ve thought a lot about why so few journalists want to do the kind of data work that drives major investigative projects or that powers immersive web presentations. The problem, I believe, is that most journalism students are self-selected to avoid anything that looks like math. At some point during their pre-college education they hit a bad math teacher, and since then have convinced themselves they must be a “word person” and not a “numbers person.” Frankly, I’m not sure it’s a good idea to have journalists who struggle to calculate percentage change operating even light data machinery like Excel spreadsheets. Back in the old days, it would take several keystrokes on a calculator to make a math mistake; thanks to computers, it’s now possible to instantly make millions of mistakes that way.

These days, students who arrive in college with good math skills and natural intuition about data are more likely to gravitate to majors in computer science or engineering. In fact, that’s where newsrooms eager to hire web developers should be recruiting. “I used to believe it was better to teach journalists to program,” says The Washington Post projects editor Greg Linch, “but in the last couple of years I’ve realized it’s easier to teach programmers journalism.”

So am I a militant anti-coder? Absolutely not. In my classes, students study journalism projects that require computer code to accomplish, and they learn where it was used. I teach them to write Excel functions that create new variables or use logical operators. I introduce them to slick data graphics tools like Google Fusion Tables and Tableau Public. They also learn basics of web page design and HTML in other classes they take at our school. Those who express real interest in going beyond spreadsheet skills are encouraged to explore coding themselves with such resources as Paul Bradshaw’s Scraping for Journalists or Learn Python the Hard Way by Zed A. Shaw.

In sum, the ability to write useful computer code is a special skill that is critically needed in modern multimedia news organizations. Journalism students need to know that such tools exist and that the ability to use them is valuable. Journalism programs with the right instructors should offer advanced data journalism courses as electives, or at least aim interested students at courses in the computer science department. But I believe that the argument that all journalists need to be coders is utopian at best and arguably unfair to the majority of students who want to develop other kinds of story-telling skills. Every newsroom needs people like me. But a newsroom filled with people like me couldn’t function.