What can journalism learn from computer science?
Journalism needs an algorithm. That’s not to say machines should replace reporters, but that reporters should be thinking more like machines: systematically. From computer programs that automate news stories, to data-driven narratives and mobile app development — journalism’s relationship with computer science is becoming ever more involved. Integrating technology into journalism, however, doesn’t simply mean installing Excel on newsroom computers, or teaching journalism students basic HTML and CSS. Applying core computing concepts to reporting and story telling can not only improve journalists’ production efficiency, but also shape their narratives.
Nick Diakopoulos, an independent researcher in computational media and former Rutgers Computing Innovation Fellow, believes journalism has much to learn from computer science. In his recent essay “Cultivating the Landscape of Innovation in Computational Journalism”, he outlined some basic similarities between the two fields. “Computer scientists and journalists both have a core interest in information”, Diakopoulos said in a telephone interview. “Journalists manipulate information on a daily basis, just like computer scientists do, but they call it ‘stories’.”
The difference between computer science and journalism for Diakopoulos is in how each handles incoming information. Computing, as defined in his paper, is “the systematic study of algorithmic processes that describe and transform information.” Diakopoulos said computer scientists look to process information, wherever possible doing so through automation, whereas journalists take a more ad hoc approach.
Diakopoulos isn’t advocating for the automation of journalism, but rather streamlining its efficiency. “Whether it’s the pen or the printing press, journalists have always been beholden to the technology that has carried their message throughout history,” he said.
In terms of careers, the computer science corner of the journalism job market is buoyant. Mark Miller, Editorial Director of Newsweek/Daily Beast, said that data journalism is one of the few areas in the industry where there are significant job opportunities. Miller returned to Newsweek after a stint at the Texas Tribune, where it had become abundantly clear to him that data journalism was a growth area. “In its first two years, data journalism drove 60% of traffic to the [Texas Tribune] site,” Miller said.
Upon returning to Newsweek in 2011, Miller was keen to harness this success and nurture the data team in his newsroom. Miller hired recent Columbia Journalism School graduate, Michael Keller, as Daily Beast’s first Senior Data Reporter. Miller said successful data journalism requires a skilled team of designers, programmers and data reporters. For Miller the core principle of data reporting is “the ability to understand what to do with the data and how to use it to tell us something we need to know.”
Through embedding journalists with strong programming skills in the newsroom, The AP is able to work on projects otherwise beyond their technical capacity. Shazna Nessa, Deputy Managing Editor of Editorial Products and Innovations said one of her top staffers — whom she described as “a journalist and incredible engineer” — was moved from the interactive team onto investigations because of their ability to create databases and find patterns in numbers.
Tristan Harris, former CEO of Apture, a startup that works with publishers to improve site retention figures which has since been bought by Google, said in a panel discussion at the 2011 South by South West conference that he sees the journalism craft being enhanced by computer science, making it more agile in its processes. Harris described his concept of object-oriented news, an idea derived from software engineering. Harris said that in the engineering world, to code twice means something wasn’t done right. Software engineering has evolved from a stage of rewriting code to adapt a process, to building fluid interfaces that sit on top of existing systems. Harris said the same should be happening in newsrooms. “We should be building on top of work that’s already done and not repeating ourselves”, he said.
Nessa, who sees a lot of resumes from recent graduates, said the best advice she can give graduating students is to have an ambient level of technical knowledge, but to specialize in one digital skill. Gabriel Dance, Interactive Editor at the Guardian US, agrees with this advice. He said graduating students should be “highly practiced in one analog and one digital skill.” Addressing the analog skill, Dance emphasized the importance of not neglecting basic reporting skills. “The best journalists”, he said, “are very good reporters”.
One simple way of putting computer science skills learned into use is to build a website. Coding a personal website from scratch not only gives students a space to practice their digital skills, they create a publishing platform for themselves. A group of journalism students in the UK took this one step further, however, and built their own live-blogging platform. Ocqur, which is still in its testing stages, is an embeddable live coverage platform built by journalists who began work on it while studying journalism at York and Central Lancaster Universities.
For Nessa, there are three core areas in which computer science can inform journalism: programming, news presentation and platform device development. Nessa also stressed the importance of asking ethical questions when dealing with data and news presentation. She said that often she sees visualizations of Twitter feeds, but no reference to the fact the data was self-selected.
In his paper, Diakopoulos detailed the symbiotic relationship between computer science and journalism. Dance agreed, saying journalism offers computer scientists the opportunity for creativity and a problem solving aspect. Diakopoulos points to specific areas of computer science his research found to be lacking in correlation with journalism. One of those areas is user-centered design; something Nessa said is an unrepresented field in the newsroom. She said design means more than simply page lay out and that more attention should be given to web aesthetics.
User-centered design — which involves focusing on end users at all stages of the design process — is one clear area requiring computer science attention by newsrooms. A current concern of many online media outlets is their website design, and an increasing number of them are adopting responsive designs to adapt for their users’ various devices. The Boston Globe was an early forayer into this area.
The Tow Center, in partnership with The Fu Foundation School of Engineering and Applied Science, coordinates a dual degree masters program in computer science and journalism at Columbia. The degree, which welcomed its first cohort of students in the fall of 2011, intends to train students in the technical aspects of digital media and news production.
Julia Hirschberg, Dean of The Engineering School, said the most exciting aspect of the dual degree is its collaborative nature. “If we didn’t offer this program, computer scientists and journalists wouldn’t be in the same room as each other.” Hirshberg said that enabling journalists to understand what is technically possible is very important to this process. “We need to find the sweet spot for what’s challenging for the computer scientists and relevant to the journalists.”
Other journalism schools are also restructuring curricula to reflect this industry shift. At The College of Journalism and Mass Communication at The University of Nebraska-Lincoln, Professor Matt Waite established a Drone Journalism Lab. Students use the lab to build drone platforms in order to learn about both the technical requirements of building drones, but also the ethical and legal issues of using these devices.
For Rutgers School of Information and Communication assistant professor, Mor Naaman, computer science and journalism is like a Venn diagram of three overlapping circles: data, information professionals and consumers. Naaman said one can “draw arrows between all three spheres” and that “every connection is mediated by computers.” Naaman, an information architect, said that before news display can be taught, a solid foundation in information literacy is paramount.
Naaman said he wants journalism students to understand computing better, including its biases and limitations. Naaman thinks it is critical for journalists to have an understanding of the core components of tech. There will be a small percentage of journalists, according to Naaman, who will be using computer science to push the envelope. “Computer science idea will be used to create new capabilities and tools for journalists,” he said. The Fu Foundation Partnership is designed to cultivate those journalists.
Nessa thinks news organizations need to take their share of the responsibility in placing students graduating with these skills on the right career path. Heads of newsrooms need to set out clear definitions of what they are looking for in their recruitment. Nessa said there should be more gradation of the types of technical skills editors are looking for because she sees a clear distinction between data-driven reporting and the post-reporting work that goes into building an interactive. The journalist on the ground, gathering facts — be that quotes or raw data — has a different workflow and skill set to the journalist who takes the numbers, crunches them and produces a visualization.
A computer science degree can benefit a journalist beyond the technical skills it teaches them. Gabriel Dance said one of the best lessons to be learned is a mindset shift. Dance, who has taught a programming course at Columbia Journalism School, isn’t comfortable with the suggestion that all journalists must learn how to code. Instead, there is another value a computer science degree — or at least elements of one — can add to their work. For example, Dance said journalists could learn a lot from logic courses. Dance has both computer science and journalism degrees and he said the biggest takeaway from his computer science degree came from his logic courses. “When you program something, everything is thought out,” he said. Dance said it is that methodical thinking that helps him when reviewing story pitches or working through information.
Emphasizing the need for fundamental knowledge, Naaman was pedantic about the way in which computer science and journalism is talked about. Naaman doesn’t like to use the phrase “computational journalism” referring instead to what he calls “computing in journalism.” A subtle but significant difference because “computational journalism” implies that the news will write itself. Naaman doesn’t believe this will be the case, he doesn’t think the role of the journalist will be replaced by a machine.
Not everyone agrees. Narrative Science is an automated news generator that uses algorithms to produce copy. Journatic, another company in the business of automating news production came under fire recently for releasing stories with phony bylines. Unlike Journatic, which outsources its news writing, Narrative Science produces stories using algorithms. It generates stories from data. Kristian Hammond, Narrative Science’s Chief Technology Officer, said “the idea is to take a journalist’s approach to data and extract the interesting parts.”
Hammond stressed that Narrative Science wants “machines to think like journalists, not the other way around.” Hammond said there are two ways of thinking about Narrative Science’s algorithms, the first being that they will render journalists obsolete; or as an opportunity for the tremendous scaling of editorial control.
Like many others working in this area, Hammond said he wants to see less of a curriculum focus on display and more on core fundamentals. Hammond said, “There are exciting opportunities in information implementation.” He used the example of search as being one are in which journalism would be able to harness computer science principles in order to editorially shape information architecture. The idea being that search would render results based on journalism principles — differing and contrasting opinions, multiple sources and historical perspective — rather than an algorithmically similar pieces of content.