Data Journalism Team Metamorphosis
As data journalism teams grow alongside their domain, the chase for the smartest and brightest talent is on — novel skills are restlessly sought after. Yet, the demand for new recruits is not homogeneous. Organizations of various sizes and shapes fight diverging battles. While new machine learning data journalists herald exciting headlines of innovation, significant future growth may loom from organizations that are yet to hire their first data journalist.
Some of the most powerful shelling in 21st-century investigative storytelling was fired from the growing arsenal of data journalism.
Stretching from one-person-shows to international renowned newsrooms, universal aspiration among news organization to turn numbers into stories have proliferated in recent years — so much so that often today’s conventional journalists pursue the more simple data journalism projects without the help of trained data experts.
As much data journalism — alias computer-assisted reporting — may have transformed the rest of the profession, it may have altered itself, even-handedly.
Signs of change are palpable - foremost, in the size of data journalism teams and genres, that its members pursue. As a victim of its own success, teams grow in size and beg for answers of how best to hire, assemble and manage data teams.
John Burn Murdoch, a senior data journalist at the Financial Times threw the question of change into sharp relief: ‘Does data journalism, like tech journalism, progress from a single specification into something that applies to more and more stories across the newsroom?’ Within the FT’s data team, the job of a data journalist transformed from once, a generalist post — an individual possessing several hats, acting as data reporter, data visualization journalists, and front-end developers, — into a multi-person operation, with stringently defined remits of roles.
That specialization of data journos follows team expansion, Paul Bradshaw, associate professor for the Master in Data Journalism at Birmingham City University, considers mere logic. With larger teams, it is natural to move from a single person having to do a variety of tasks, to a more allocated team effort: “Visualisation gets delegated to the designer, coding to the developer, writing to the journalist. As the team grows, the need for an extra person to manage those projects, grow too. [As] ambition grows — they may identify strategic areas to recruit in, such as machine learning”.
Novel innovation in this area is versatile — spanning now from drone journalism, machine learning journalism, Newsgames or technical upgrades to more basic data crunching.
Training a computer model to confer news ‘with an edge’, recently unimaginable to be deployed within punier newsrooms — L.A. Times, BuzzFeed News, The New York Times — all seemingly, in one way or another, gave machine learning a chance within the newsroom.
Those who couldn’t work it out so far, may now opt for hiring new expertise or developing skills internally. Those who can’t — notably smaller news organization — may choose to partner and share expertise with others, a conceivable workaround, as the example of Quartz shows. The seven-year-old news website announced late last year plans to utilize machine-learning methods to ‘publish stories that would otherwise be impossible’. The difference: Not entirely self-benefiting. The upshot: It aims to collaborate with other news organizations, ‘for at least half of their published stories’.
Hiring talent means making choices
Marie-Louise Timcke, heading the interactive team at Funke Media in Berlin, says that assembling a data journalism team would largely depend on how one answers the following questions: “Do you want to reach a maximum number of people? Do you focus on long, engaging stories? How visual, how investigative do you want them to be?’
Timcke also agrees that with increasing team size, specialization would increasingly mature. However, despite her relatively small team of five, specialization already surfaces — there’s one with a journalistic background, a UX/UI designer, two programmers, and herself — a hybrid between journalism and statistics.
As Funke Interaktiv occupies the data story assembly line with data journalism specialists, it was impossible to avoid overlaps responsibilities roles — such as within areas where expert knowledge remains scarce and everyone would need to chip in to figure it out.
Timcke also speaks of a perfect data team size — of 3 to 7 people — for a smaller data newsroom that produces a whole project from start to finish, in one pull. If the team would grow beyond that, one would have to juggle several projects at the same time, which, in turn, could make the magic of a single-team-effort fade.
Germany’s data journalism appears to be a level playing field. Different newsrooms cover different remits. The team at SZ cover more investigative topics, and the team at Spiegel cover daily data journalism projects. Funke pursues more of the visual and interactive kind of stories. It might be conceivable that this setup makes it less attractive for larger newsrooms with more resources to simply poach talent from smaller operations — as talent would need to be retrained, but clear evidence beyond spotty anecdotes does not exist.
The evolution of data journalism teams and its talent may be especially instrumental for investigative newsrooms, many of them small. Jennifer LaFleur, data journalist-in-residence at the School of Communication at American University, thinks that new skills have emerged to allow journalists to produce stories that were not conceivable in the past.
Caelainn Barr, Data Projects Editor at the Guardian in London — who runs a team of three data journalists — says that she encourages her data staff to spend time on developing their own skillsets. This might occupy as much as one-fifth of their time, but the rewards make it worth it. The narrative behind Barr’s strategy? She wants to push more investigative stories — many times investigations that other journalists, at first, may not immediately identify as data stories — to desks and reporters (instead of merely receiving and answering data-driven inquiries).
Coverage across the newsroom landscape shows that new angles can emerge from, for instance, satellite imaging, machine learning and the processing of large data dumps and piles of scanned pdf documents. Data journalists in smaller, resource-constrained newsrooms can face challenges to manage this all on their own. Forming and managing internal alliances with other members of newsroom staff would become more vital. “With maybe only one person dedicated to data journalism — and that is when they are lucky”, adds LaFleur — “collaborations are key”. Projects like ProPublica’s local journalism initiative would add to its merit by helping to provide readers with their fair share of local investigations and data-driven reporting.
Does supply of new data journalists match demand?
Adrian Blanco, presently enrolled at one of US East Coast’s finest journalism programs, geared towards data journalism, at Columbia Journalism School — speaks from first-hand experience when attesting that size usually matters when newsrooms opt for generalists or specialists.
Primarily, US-based mid-sized newsrooms are eminently choosy in their recruitment selection: “They might specifically hire for someone who makes maps or creates d3 visualizations”, he says. On the contrary, smaller newsrooms would work upon an entirely different rulebook. Teams (such as the Los Angeles Times, Chicago Tribune, Miami Herald or The Atlantic ) would generally shop more broadly for data journalists as the individuals tend to operate across departments and desks. As a result, the appetite would be more geared towards the ‘jack-of-all-trades’ type. (Larger newsroom data journos tend to work in specific designated teams or departments — perhaps in a graphics or statistics team).
Perception of the data journalist
The standing of the data journalist in newsrooms have also shifted, observed Mr Bradshaw. From a specialist position once occupied by more senior, experienced journalists — who added data to their existing skillset — it would have morphed into a position now often entry-level and occupied by more junior journalists. The setup can provoke a sense of disempowerment among data reporters and developers when arguing in favor of new, maybe different approaches, when senior non-data editors, have their final say. Another peril is that a clear career path for the data journo may be absent or at a mere infant stage, often about to be developed, which may add to a feeling of disillusionment. Finally, there is also an issue related to salary expectation. Data journo developers are usually underpaid. How do you keep data journalism engineering talent when non-journalism roles are paying so much more competitively?
But despite lower salary, enthusiasm among wannabes remains upbeat. With brand new batches of talented young data journalist graduating every year, supply seems guaranteed. But ‘demand would not follow supply’, Bradshaw insists— who has witnessed within his own University walls the constant nagging by employers who keep asking for students to fill skill gaps.
In respect to the emerging crop of new data journalists, Blanco says it is important to bet on specialization if aiming at differentiating yourself and heading for a medium to large-sized newsrooms. But a sound strategy for the job market is to have a general base level in all data journalism disciplines — scraping, data gathering, data analysis, visualization, and reporting.
As teams at global newsrooms keep expanding, a priority would be to plug specific skill gaps by scouting and hiring designated expertise (as opposed to adding all-rounders to teams). Columbia School of Journalism caters to this need. Blanco’s program encourages its data journalism students to foster specialization in one area of choice — whether mastering wizardry with news-bots or other novel techniques relating to reporting.
“The end goal is to have data journalists, who can build up their technical expertise, to a level where they can do most things’,
Data Journalist @BBCNewsGraphics
Another strategy to ensure supply of new data expertise and maintain know-how within news teams is to grow it organically, says Daniel Dunford, who started as data journalist at the BBC three years ago: ‘As we got bigger, to know coding language grew in importance. When I joined, advanced use of Excel was the only data analysis tool I was using. Now everyone in the department uses R, some use Python’.
For the seeds to prosper within a generalist team, a knowledge-sharing culture is essential: “We got quite a good sense that if someone learns something, it is always shared around the whole team. The end goal is to have data journalists, who can build up their technical expertise, to a level where they can do most things’.
Since Dunford’s joining, his team doubled in size. Data journalists would predominantly be assigned in accordance with their knowledge and preference for a certain beat — not by their technical skill level as it is the case in many other newsrooms: ”A politics story, there are a couple of people this story would go to. If it is a health story, it would go to someone else.” It raises an important question of whether the future of data journalism will establish data roles closer to newsroom desks (as opposed to having their very separate unit or corner).
Will prospective data journo job posts increasingly advertise for health, business or finance data journalists — as opposed to a machine learning or python data journalist? It appears conceivable. It might strongly depend on the newsroom culture and how data journalism is treaded at present.
The AI data journalist
As mentioned, techniques such as machine learning appear to have its place somewhere in journalism, but ascertaining the true merits of it at a large scale and justifying carving out and arguing in favor for budgeting for new roles doing only AI work in a news related context, remains tricky.
More direct and significant growth could derive at a more basic level — potentially, in areas, where the merits of data journalism work is largely confirmed. Bradshaw identifies major vicinity of growth to be two-fold — a larger part not even coming from [data journalism] teams, but rather from organizations that are about to acquire their first data journalist. Moreover, growth could be powered by non-data journalists with data journalism skills, using those in their work.
More women in data teams
Do women have a fair chance in the profession — both as individuals and as part of data teams? Sarah Newey, Reporter, and Data Journalist Global Health Security at The Telegraph in London says that in her career so far she only encountered teams that are relatively fairly distributed.
Yet, she prescribes to “ensure that women feel empowered and have role models (for instance, Mona Chalabi, the celebrity of data journalism)”. Fair handouts of fellowships to women also appear crucial. Sarah herself entered the area through a Google News Lab Fellowship after her internship at her paper drew to a close. From what she knows and hears from friends in the field, she concludes that “increasingly women are represented and — even if change is slow — it is definitely in the right direction”. There are other signs of hope. In the past years, several women emerged as distinguished voices within the data journalism industry — some, in the form of members of the prestigious Data Journalism Awards jury. Esra Dogramaci at Deutsche Welle was quoted to have said that women, in general, were underrepresented, not just in news coverage but in leadership positions. Her own employer, she says, offers a much more egalitarian working environment.
Caelainn Barr adds that there are so many talented female data editors and journalists out there that there is no excuse not to have equal representation of women in data journalism teams. She recommends: “If women are not fairly represented on a team you have to ask yourself as a leader and a company how are we failing to attract and retain talented women in this field and what can we do better.”
Techjournalist is an investigative journalist interested in tech, emerging markets such as China and Africa, finance, data models and statistics. He has a background in business and economics, coding, and has a passion for the application of data science within the newsroom. He lives with his wife in London.
