Interactive-Journalism-Chapter 2.pdf

2 The Rise of a Subspecialty Interactive Journalism NPR’s media columnist David Folkenflik delivered a breezy report in 2005 from Lawrence, Kansas, home of the University of Kansas. He nar- rated his lead-in: So here’s The World Company at work. It’s got a daily newspaper, Th...

2 The Rise of a Subspecialty Interactive Journalism NPR’s media columnist David Folkenflik delivered a breezy report in 2005 from Lawrence, Kansas, home of the University of Kansas. He nar- rated his lead-in: So here’s The World Company at work. It’s got a daily newspaper, The Journal- World, a local cable TV channel, and, of course, several Web sites, including one with video games. That’s right, one of the Web sites goes to the trouble of staging an Xbox version of every University of Kansas men’s basketball and football game. Video game replays of every basketball game in the Jayhawks’ history. This was interactive journalism at work. And this was three years into the online experimentation at The Lawrence Journal-World. By the mid-2000s, the news- paper with circulation of a mere twenty thousand had caught the attention of the national media, such as NPR, and big media companies were drawing inspiration from The Journal-World’s innovations, including the likes of The New York Times. So many news organizations were visiting and taking tours in order to learn about online innovation that, as a Chicago Tribune article described, “the interest grew so intense and so disruptive to the workday routine that The Journal-World had to schedule visiting hours.”1 The Lawrence Journal-World had been busily doing interactive journal- ism, with staff journalists who could program creating new features for the Web. For example, as early as 2002, a major news effort intended to examine drought in the state was accompanied by a vibrant interactive experience. As Rob Curley, head of the Web site in Lawrence, explained to Folkenflik:  We put together a map of the state of Kansas where you could mouse over any county in the state, and we would show you what their normal rainfall total would be and what their current rainfall total would be and what that meant in millions of dollars to the local farmers. Then we had our nerds work with the nerds at the state of Kansas, and we built a database of every well in the state of Kansas. So you could see how irrigation was affecting the well water supplies.2 The Lawrence Journal-World was not the first place or the first time in the brief history of online news that interactives were created, but the special combination of people and the types of projects that emerged solidified the scene as a defining moment in the history of interactive journalism. For the first time, journalists in the newsroom were actively using programming language to render data and information in visible and interactive ways, not just across isolated projects, but woven into the core fabric of the Web site. Adrian Holovaty, a former staffer at The Lawrence Journal-World and now a recalcitrant developer of music software after deciding to quit working in journalism, was at the time a programmer-journalist wunderkind who helped solidify a small community of other journalists interested in the possibilities of programming and journalism on his blog and who became a model of the future of the subspecialty. What happened in Lawrence was nothing short of remarkable and prescient, and it is for this reason that Rob Curley, now the executive editor of The Orange County Register, doesn’t go a single day without thinking about Lawrence. He explained to me, “We were building things that people really wanted; people had never seen Internet behave this way before, and it just worked really well.” Almost no one had seen an interactive in 2002. In fact, a bare majority of Americans had the internet in their homes, and the vast majority of those only had slow dial-up service. The modern history of interactives doesn’t begin with Lawrence, but it is the most dramatic example of the kind of innovation that would come. In the rest of this chapter, I’ll outline some of the history of interactives based on interviews with key figures, as well as discuss key technological developments I’ve identified over the course of my research and reading. This history of interactives is not intended to be exhaustive; the goal, rather, is to underscore the rise of interactive journalism so we can understand how a subfield comes to be. The Early Rise of Interactives Inasmuch as interactive journalism can be attributed to the rise of sophis- tication in code and computing technology, and draws power from the vast amounts of data present today as never before, it nonetheless has a long his- 38 Chapter 2 tory and evolution in news. The rise of the subspecialty emerges from a fusion of photography, graphics, data visualization, and computational journalism. These aspects of journalism all intersect in different ways, merging together to help form interactive journalism. The form of news itself—its presentation—has evolved throughout history, and the study of it yields some insight into the rise of interactive journal- ism. Kevin Barnhurst and John Nerone chronicled some of these changes, arguing that the persistent visible structure that repeats day after day, which includes layout, typography, and illustration, presents the look and feel of newspaper, which in turn structures the relationship between it and the public.3 They categorize newspaper presentation into four formats: printerly, partisan, Victorian, and modern. The printerly newspaper, with its bookish appearance, was for the gentleman; the partisan newspaper had a larger format and suggested the rise of mass politics and a market economy; the Victorian newspaper reflected imperialism; the modern newspaper expressed bureaucratic production and offered expert explanation. Interactive news today builds on this legacy, offering user-guided experiences of news and information that may structure the relationship between press and public as demanded by the previously outlined pressures on journalism today. The history of images in journalism, particularly graphics, also contributes to the rise of interactives. Mark Monmonier, whose Maps with the News represents the most comprehensive account of graphics in news, argues that the beginnings of images in news can be seen as early as the arrival of block printing in the sixth century, where raised images were produced by a roller.4 John Grimwade, a former graphics editor at The Times of London and then Condé Nast Traveler between the 1960s to the early aughts, is an amateur historian of infographics in the news. He suggests that the earliest infograph- ics appeared in the early 1800s and mainly focused on battles and murders. According to Grimwade, the first news infographic was published in 1804 in The London Times and showed a house that depicted a murder scene as the crime progressed. “It was one of the first narrative news graphics. It a pretty straightforward idea with the rooms of house showing where different things occurred,” he recalled.5 Charles Minard’s graphic depicting Napoleon’s march to Moscow is often incorrectly credited as the first news infographic, as it was produced in 1859 but depicted an event that took place in 1812. The first weather map didn’t appear in news until the end of the nineteenth century. Other major leaps in infographic expression, though, did not happen until the twentieth cen- tury, when the sinking of the Titanic and World Wars I and II offered ample opportunity (and desire from readers) to use infographics to depict ongo- The Rise of a Subspecialty 39 ing news events. Time magazine was considered the leading source of news infographics during this era. Grimwade argues that advances in information technology—not necessarily the technology to make infographics, but the speed to communicate information and better data—improved infographics greatly (for instance, real weather data). Graphics specialists were essentially like draftsmen and had technical drawing skills. They used French curves, ellipse templates, compasses, air- brushes, knives, and, of course, pens. In Grimwade’s time (the early 1970s), to trace lines, these graphic journalists had to cut down on acetate plates to prepare for publication. He reminded me that there was no easy way just to fix a graphic—no computer “erasing” to adjust angle or correct an error. A graphic journalist had to have everything essentially perfect. As Monmonier tells it, the 1970s and 1980s saw many innovations in newspaper design that actually led to a more prominent role for photographs and graphics. He argues that graphics accelerated in part due to technological factors and in part due to competitive factors. By 1979, with the founding of the Society for News Design, the field had become a subprofession of its own. Part of the rise of the graphics department as a distinct specialty within journalism came with the ability of newspapers to use color. As technology improved, both in printing presses and layout, newsroom executives began to look more favorably on the possibility of incorporating more sophisticated graphics in news. Broadcast journalism had begun incorporating graphics in news much earlier (with the advent of TV weather forecasting), and print newspapers began to catch up. Monmonier argues that in the 1970s and 1980s, newspapers began to form full-on art departments that would make the newspaper “attractive and informative”—such that there would be a variety of illustrations: maps, editorial cartoons, business graphics, color photos, and illustrated layouts for special sections (Food or Travel, for example). Early computer graphics came into the newsroom in the mid-1980s. Mon- monier noted that “the programmer and the artist collaborated,”6 producing charts and graphics. The mouse-controlled arrow allowed for easier online creation, and software (Microsoft’s Chart, for example) made it easier to make hectographs, time series graphics, and data charts. And the newspaper’s editorial front-end word processing became linked to the typesetting, mak- ing it easier for computers to transmit graphics. The Associated Press began moving graphics over its electronic bulletin board, and Knight Ridder and Gannett put Macintoshes for graphic design into newsrooms. In addition, graphics departments were even more formalized according to newspaper structure. He notes that in 1984, graphics officially became a specific unit in the AP under its own editor. And some of those who came to the graphics 40 Chapter 2 desk had backgrounds in art rather than journalism. USA Today’s found- ing in 1982, though controversial, elevated the prominence of graphics as a significant aspect of the marketing of journalism. Monmonier leaves off in 1989, but he makes some cogent predictions about the state of interactive graphics that he believes will emerge in journalism. He writes about the promise of videotext for interactive capacities that will allow user-directed control over information: Electronic display makes possible map symbols that blink and move, and com- puter graphics systems can generate animated maps. Video display technology and modern telecommunications offer the potential for linking highly dynamic cartographic display software with large geographic databases.... Should elec- tronic news databases ever emerge as a mass-communications medium, dynamic news maps would become commonplace.7 In turn, a dynamic map would give users a “menu of choice” to make selec- tions through a keypad—or even a touchpad, he predicts—whereby computer coding would offer promising ways to focus attention of the user on new kinds of information. From a practical and philosophical perspective, the ideas of interactive Web design were predicted as a fundamental way to en- gage the computer user of the future, and, for news, the way forward would be user direction through electronic news. In news, the rise of multimedia capacities and the improvement in Web sites led to early innovations in interactives by the 1990s. Michael Friendly argues that Fortran in the 1970s ushered in highly interactive capacities for the creation of interactive graphics,8 but it is only in the mid-1990s that we start to see Web-based presentation of interactives in the news. Friendly points out that advances in computing languages made it possible to develop sliders, selection boxes, lists, graphics and tables, and user communication through messages. Interactive graphics in news may have been born in 1995 at The Chicago Tribune, though it is hard to for any one paper to claim to be first. But a conversation at The South Florida Sun-Sentinel is indicative of the innovation taking place at the time. In 1996, Sun-Sentinel editor Leavett Biles recruited Don Wittekind, a journalist who had some programming background, to help create what Biles called “multimedia informational graphics”9—though Wittekind said he had no idea what they were. Their first effort was a graphic on fire ants, which interactive news journalist Alberto Cairo described as “uncharted territory of play buttons and moving around rather than plan old ink, paper, and still pictures.”10 At this point, these newsrooms were experi- menting with the beginnings of human computer interaction—with feedback The Rise of a Subspecialty 41 between readers and computers soliciting an actual reaction, be it through visuals, sound, or an onscreen response.11 So, what was it like, then, at the dawn of the electronic, computer-enhanced interactives? The next section offers some vignettes from the era. The Start of Modern Computer Interactives In the late 1990s, newspapers and news organizations more generally came to the overarching and seemingly obvious conclusion that they needed to invest in robust Web sites. Major newspapers had been burned by early attempts at innovation. As you’ll recall, Knight Ridder and The New York Times, for example, had invested in Videotext technology, a precursor of the modern personal computer with information transmission facilitated by a modem. But home adoption had been slow, and there were few commercial returns on the service. Pablo Boczkowski provides perhaps the best chronicle of how the print daily newspaper came online. According to his research, economic motivations spurred investment in the Web; as consumers moved to nonprint alternatives, newspapers were already beginning to see declines in profit. And by the end of the 1990s, more than 40 percent of the adult population was online.12 Andrew DeVigal, now a professor at the University of Oregon but a for- mer longtime Chicago Tribune and New York Times interactive journalist, described to me the early days of interactives. He was part of the team that built the first interactive at The Chicago Tribune, if not one of the first inter- actives ever, and he also helped build the newspaper’s first Web site. “It was six months from start to finish before we launched. It was one hell of a ride setting up templates. There were no technologies that developed templates for the first year or so [1993–1994]; The Chicago Tribune had to do it by hand.” He explained that there was “this staff of about twelve young people in this room on the fifth floor, and people in the newsroom called it the ‘dorm room’... there were all these kids in this room putting together this thing, with music blasting and it was as fun as hell and it was sort of renegade.”13 At first, The Chicago Tribune was only experimenting with video. But for what the Web was capable of in the early to mid-1990s, their video efforts were unsatisfying—they came out small on the page (150 pixels, about 1.5 inches) and there was no professional staff devoted to video work. But a major breakthrough came with what DeVigal claims was the first news interactive, though this is hard to confirm. In 1995, The Tribune launched an interactive that tracked every homicide in Chicago, a project that continues in a different form through to today.14 The team working on the project had to figure out 42 Chapter 2 how to make the map interactive without having any obvious technology tools to do so if they were to achieve their goals of enabling people to look up homicides by zip code. They relied on CGI, or the Common Gateway Interface, which was an early (though still used) script that enables a user to send a request to a Web site and retrieve the output (a pre-programmed map, for instance) without the output having to be shown on the Web page. Essentially, this script was a way to create a sense of a dynamic Web page, even though every request for a homicide had been preprogrammed and hand coded. The team literally had to type in zip codes and build the graphics GIF by GIF—all by hand. This kind of tremendous effort would become seemingly effortless (with no hand coding necessary) as programming and the Web more generally evolved. There were signs that the specialty of interactive journalism was beginning, even in these early days. DeVigal and his staff enjoyed the status of having special skills—including the ability to program and use programs—that no one in the newsroom was using. Their mission was different, too: to tell stories in new ways using new technology. This special mission was beginning to be articulated by higher-ups and the teams in newsrooms in Chicago and elsewhere. DeVigal emphasized to me that interactive storytelling was interesting to The Tribune, to him, and to others, then and now, because it gives users a chance to “truly control the power of their interaction and gives a choice of how to experience a story,” a theme we will see elsewhere. Interactive journalists began to develop a specific way of thinking about journalism that would later quite possibly change journalism norms, or at least offer disruptive ways of thinking. An occupational identity started to come together during this period as well. DeVigal participated in an early panel on interactives in 1995 at the COMDEX conference in Chicago. The conference is now no longer in ex- istence, but it was known as “geek week” and at the time was the one of the largest computer trade shows in the world.15 There, DeVigal began meeting others working on similar projects related to interactive storytelling, learning about new techniques and projects that, as he recalls, “... blew my mind. I saw people doing things I want to do.” A professional network was begin- ning to coalesce, one that would later emerge as a key sign that interactive journalism had come into its own, as it could be measured by internal and external signs of influence, including the number of practitioners. The field, though, was still disconnected, and there was little isomor- phism between newsrooms about tactics, strategies, and forms for inter- actives—unlike now, as interactives tend to have the same underpinning normative approach to journalism and, despite varieties in subject matter The Rise of a Subspecialty 43 and presentation, have much consistency across newsrooms. But one factor of interactive storytelling was clear: it was laborious no matter where you went, and it required specialized skills and the ability to “hack” within the confines of existing Web and programming limitations in order to get the projects to work. Elsewhere, building interactives was just as laborious. Geoff McGhee, an early pioneer in interactives who would go on to lead The New York Times’ first iteration of its interactive team (the online graphics desk), began rudi- mentary efforts starting in 1997 and into the early 2000s. At the time, though, ABC News, McGhee’s employer, had a reputation for innovative Web design and a robust team fueled by significant investment from the network. He recalls that it “took forever” to download a single image on the page, and monitors were at the time still tiny—eleven to twelve inches. With the size of these monitors, and the fact that the predominant browsers of the time, Netscape and Internet Explorer, had so many rows of search buttons and marketing offerings, visual presentation was difficult.16 At ABC, McGhee couldn’t figure out how to create an interactive with a unique URL. He explained that the only way he could enable an interactive was to code JavaScript in a way that created a pop-up window (like a pop- up ad today) that made it possible to put images on top of each other on the same page. He created maps by putting transparent gifs on top of each other, perhaps first a map, then numbers, then text, and it was possible to click on the numbers or text so they would reveal still further images. “It did give you the sense that you were clicking on an interactive, it was a way to deal with things visually and get a story that went beyond search and scroll,” he said. McGhee explained that these “magical pop-up windows with no links” still “really did go beyond episodic journalism.” He and his team created polls, quizzes, backgrounders, and a specific project, he remembers, that got significant attention: individual pages for each Supreme Court justice with a drop-down menu that would allow a user to see the Supreme Court docket and further information. Although the interactives weren’t fancy, these developments were solidifying how interactives could provide alterna- tive ways of presenting the news and telling stories to people. DeVigal and McGhee were both journalists who were blending their journalism skills with programming. But other newsrooms were doing interactive experiments, challenging perceptions of storytelling, and redefining occupational identity, but they had not quite realized the vision of a programmer journalist. Boczkowski offers insight into how journalists were also having trouble with programmers, as many journalists interested in interactives had not quite become programmers and programmers had not quite embraced jour- 44 Chapter 2 nalism. He recalls this in a section of Digitizing the News that examines The Houston Chronicle’s online division.17 From the late 1990s to 2000, a team of journalists, paired with a programming team, worked on an ambitious project called Virtual Voyager. The effort began when a businessman and a nurse contacted the newspaper to see if it had interest in covering their three-year circumnavigation of the earth. The print travel section passed on the project, but the online division was interested. “At Sea” came to log over one thousand entries and included photos, audio, videos, and a map that automatically updated with the voyagers’ location.18 This was but one Voyager project. Boczkowski, too, argues that this effort went beyond just episodic journalism, telling a story in a different way. Programmers worked apart from the editorial staff and had demands beyond just working on the Voyager site. While eventually the Voyager and design/programming staff began working together well, the project was more a process of translation between editorial and technical, requiring the coordination of a variety of occupational identities: editorial, programmers, designers, and production staff.19 Journalists weren’t programming yet; the head of the project observed that there were cooperation difficulties between programmers and journalists that could have been solved if he had known more about programming: “Part of it is my fault. I should learn Java.”20 There were other precursors that suggested both the differences inherent in interactive storytelling and the formation of a new professional identity. As that same journalist remarked, “In print journalism, we wanted to go out and experience something and then come back and put it on a sheet of paper. We didn’t notice the movement [or] hear the sound except in terms of something we could translate into a printed product. I think Virtual Voyager is making us open our eyes to a form of journalism we didn’t need when we were print people.”21 Here, we see the development of a consciousness of both the difference of interactives as well as the proto-formation of an occupational identity distinct from the “print people.” This kind of recogni- tion—as being able to provide different contributions—was critical to the rise of interactive journalism. Other early examples of projects that were less visible to those outside the newsroom were the interactives that were being built as tools just for jour- nalists to do their work—databases, searchable maps, and other versions of interactives for internal rather than external audiences.22 In the early 2000s, Tom Torok of The New York Times (and formerly of The Philadelphia Inquirer) demonstrated a tool at a NICAR conference that, as ProPublica’s Derek Willis writes, “would make your SQL Server database searchable. Any database, no matter the type of information contained in it.” He called it “Shboom.” The Rise of a Subspecialty 45 Willis open-sourced the code for Shboom in 2014 and, in this accompanying post on the online repository GitHub, also noted, “It was a remarkable piece of software for those of us who had no idea what software really was. It is one of the first examples of web development in the newsroom that spread beyond it.”23 Developments in Europe may have begun slightly later, as most of the jour- nalists I spoke with there pointed to the development of Macromedia’s Flash (around 1996) as critical to beginning their interactive career. In Spain, at both El Pais and El Mundo, journalists began transforming infographics into interactives. Alberto Cairo recalls that the September 11, 2001 attacks were what convinced him (and perhaps the rest of the newsroom at El Mundo) that interactives were going to be part of the next wave of the development of online journalism—and certainly the next iteration of infographics. In the United Kingdom, the BBC was an early pioneer of interactives, launch- ing initial efforts in 2000 and 2001. But, as a number of the journalists I spoke with and Boczkowski also suggest, many early adopters were “print people” adapting to online contexts and innovating from a print perspective. A more fluent, online-focused and/or computer-literate journalist would later dominate the specialty. Nonetheless, these early days set in motion what would come to be a way of thinking about interactives that would coalesce industrywide and help lay the groundwork for establishing a professional occupational identity as distinct from the rest of the newsroom. Lawrence’s Many Contributions The online team at The Lawrence Journal-World worked in the basement of the newspaper building, which was itself located in a rehabbed old Post Office building. Instead of the subterranean space being dingy or smelling musty, the floors were a warm, inviting wood, and visitors were surrounded by red brick walls. Rob Curley described the space as “gorgeous, as if you were walking into the coolest bar and grill you had ever seen.” His team crouched around Macs and, thanks to the wide latitude given by the Simons family, benevolent owners, and a flexible print staff, Curley’s online group was given the opportunity to innovate. Projects included a tracker for the state legislature, a database for profes- sors’ salaries at KU, a database of every statistic in every box score for every KU football game since the 1890s, a statistical database for every high school player in Kansas, automatically updated weather maps, bar/beer/restaurant specials and guides, interactive and searchable music listings, and beyond. A major innovation was thinking about journalism as structured data— 46 Chapter 2 Adrian Holovaty later wrote an essay about the idea, where he explained, “An obituary is about a person, involves dates and funeral homes. A wedding announcement is about a couple, with a wedding date, engagement date, bride hometown, groom hometown and various other happy, flowery pieces of information.”24 In other words, journalism was made of bits of component pieces, and with the right programming, people could search of their own accord to find the pieces they wanted. “Adrian opened my eyes to the idea of a relational database,” Curley ex- plained. “We didn’t know what you should do or not do. We thought, wouldn’t it be cool if you had a database with drink specials in Lawrence? Yeah, that’s totally cool, so let’s build that. How about let’s build a bar database with a calendar of events? What about layering photos? What about making all shows searchable and then having downloadable mp3s of the band?” Curley described a typical Lawrence project. The University of Kansas an- nounced in 2003 that it would no longer honor alumni season tickets as it had in the past, assigning instead a lottery according to a different set of criteria. Reporters got ahold of the new criteria, which included alumni status, giving rates, and length of time the seats were held, among other factors. The online team created an interactive map of the stadium along with a database that would enable users to sort where their new seats might fall based on the data they entered. They could then click on the predicted seat and get a view of the stadium. This sounds perhaps unremarkable for today, but the Lawrence team combined programming, data, and photos and then built interactives in step with traditional reporting to give people a chance to experience a story in a way that was personal to them. Nothing like this had been done before. Other Lawrence projects included innovations in election coverage. The Lawrence Journal-World hosted live online chats with candidates. Holovaty described on his blog an early interactive: “[We made] interactive ‘candidate selectors’ for the city-commission race and school board race—click on the candidate quotes you agree with, and the script will tell you which candidates you agreed with the most. It’s intended mostly as a guide, but it is, without a doubt, useful in that ‘Hey, that really made me think about the candidates’ sort of way.” Here, election information was being retold in ways that gave readers a direct, user-navigated experience of learning about candidates. This kind of approach has become a hallmark of interactive journalism. In addition to creating interactives, Holovaty and another staffer, Simon Willison, a programmer recruited from England, also built a Web framework, Django. The goal of Django was to give Python programmers a starting point to build Web sites and Web features that were often used by news organiza- tions, but Django was and still is not just for news. However, the two did The Rise of a Subspecialty 47 develop a content-management system bolstered by Django for newsrooms, called Ellington, one still used in newsrooms today. The blog Holovaty kept at the time became a focal point for the small community of journalists like him and journalists who were interested in what online journalism could create, particularly when programming and journalism were combined. On it, he wrote about experiments in Lawrence (and elsewhere as his career continued), critiqued news Web sites about almost everything from user registration practices to the difficulty of find- ing someone to email on the homepage, and he helped build community through posting job ads from around the country, including at some tiny newspapers along the “Treasure Coast” of Florida. The comments created a lively community that fostered sharing of experiments and ideas—those involved would go on to have significant roles at major institutions, includ- ing The New York Times, The Washington Post, NPR, Facebook, Twitter, and The Advocate/Here Media, among other media and tech companies. As we saw, before there was Holovaty, there were others laboring in pock- ets of newsrooms across the country and the world, experimenting with interactives as ways to tell new types of stories. But Holovaty became a focal point for the myth of the programmer journalist and would see it to his peak involvement in journalism and programming in 2007. Lawrence was part of the myth of the birth of this programming-journalism convergence; but it’s important to give credit to others working in this area—especially those who saw the promise and importance of developing a new subfield of work. And technological developments (created, of course, by people) made a big difference, too. Technological Developments Technological developments contributed significantly in enabling interac- tive journalism to play the role it does in newsrooms today. These inno- vations empowered journalists to create better and better projects, and as the technology for consuming them also improved, the contributions that interactives could bring to the news experience became clearer. Similarly, these advances are also not necessarily linear: one development does not, as a matter of course, flow into or set the conditions for the next—but the way journalists have described these changes is constructed as a holistic narrative that brings interactives from linkless pop-up windows to sophisticated, per- sonalized, even virtual-reality-enhanced projects—and further emphasizes the special knowledge and contributions interactive journalists make to the profession. The technological developments include the increasing sophisti- cation of the Web itself, the rise (and decline) of Flash, the importance of the 48 Chapter 2 rise of the “social” open-source community, and cloud computing. Some of these innovations are technical and jargon-y, but I hope, by reviewing what they have contributed, readers can acquire an appreciation for how things look today—and why interactive journalists have a special claim to esoteric knowledge, both practical and normative. Table Percentages and Iframes In the days of early interactives—most Web pages were constructed of HTML (HyperText Markup Language), which was complicated for the time but would seem simple now. HTML was well understood, although it was lim- ited. The introduction of tables allowed spreadsheet layouts to be put into Web pages, and it soon became clear that it was easy enough to get rid of the visible lines and construct column layouts. Tables enable programmers to create more visually interesting pages and facilitate the placement of captions, titles, navigation bars, sidebars, and the like. A major innovation, according to DeVigal, was the ability to play with table percentages, or have more con- trol over the sizing of these elements, and thus it was easier to create more customized and better-looking pages. Table percentages enabled designers to do with Web pages what someone might do with a traditional print layout. As mentioned, the rise of CGI was important because it became a powerful way to experience a more dynamic Web. It would take input from a Web page and then run a program to generate a new Web page. It has limitations because it is slow—it requires recall to a server—rather than other develop- ments (for example, asynchronous Javascript or Flash) that engage with the local computer rather than the server. The table developments were followed by the rise of iframes, which further enabled interactives. An iframe makes it possible to embed a Web page within a Web page. Think, for instance, of an image on a page that can have its own scrollbar, or other configurations that can be changed without navigating or reloading from the hosting page: only the content of the page changes, not the entirety of the page itself. This technology meant reduced loading time, which, in an era of dial-up modems, was everything. Iframes further enhanced the ability to build interactives that offered a better user experience and featured more dynamic properties. Flash Interactive journalism was taken to a new level when Adobe’s Flash was invented. By the mid 2000s, Flash became integrated in newsrooms as the critical tool for making interactives. Flash has its origins in a program called The Rise of a Subspecialty 49 Macromedia Director, which was used to make multimedia content. Shock- wave, a plugin, was used to play Director files within Web pages. Adobe bought Macromedia in 2005, and from this foundation created Flash. Thanks to Flash, it became relatively easier for people who did not know program- ming languages to build interactives, and the software had a short learning curve. The New York Times had been an early experimenter in Flash begin- ning in 2004, producing political interactives such as election maps.25 By the mid-2000s, Flash, based on the ActionScript programming language, enabled easier video hosting, dynamic displays of visualizations, smoother integration of photos and audio, and was well used and well loved by many newsrooms and even tech companies (when YouTube began, it relied on Flash for its video player). Flash had great traction, in part because the Flash plugin was bundled so well with browsers, and since every browser had or could have the plugin, almost every user could access content created by Flash. Flash also aided the transition of infographic specialists who had worked mainly in print design and brought them online, as they were able to pro- duce their products with the same kind of visual details as print but with enhanced capabilities online. Some of these infographic specialists (Cairo among them, as well as Jon Keegan at The Wall Street Journal, whose team is profiled in chapter 4, and others) have emerged as thought leaders and key figures in interactive journalism. Flash played a major role in popularizing interactives because they were easier to make, so journalists could make more of them—underscoring interactives’ value to the newsroom—and their quality helped make a case for their relevance as storytelling vehicles. And Flash also enabled more would-be interactive journalists to become part of the growing subfield. However, Apple essentially killed Flash on mobile devices by not allow- ing it to function on the iPad and iPhone starting in 2010.26 What had made Flash so appealing—its cross-browser integration—was no longer available. I distinctly remember in 2010 (after the iPad was first announced) when New York Times tech writer Nick Bilton was coming up with conspiracy theories about the demise of Adobe due to Apple’s decision not to allow Flash on the device. Given the rise of mobile, this meant that if newsrooms continued to create interactives in Flash, they wouldn’t work on the most popular mobile devices. Though some interactive teams were already programming outside of Flash, the forced end to Flash meant that people who knew programming languages became hot commodities in newsrooms. If you could program an interactive in Java or Ajax rather than Flash, you were wanted. As Cairo describes, “Using Flash was so comfortable and so powerful, and you could create some very sophisticated work, not with 50 Chapter 2 minimum effort but with reasonable effort. Why would you care about learn- ing about JavaScript or programming languages [when] you could just leave that to the developers? But once Flash was gone, we were in the desert and we saw we needed to embrace other technologies used by people who were not journalists.” For example, in 2009 The New York Times was still using Flash to create sophisticated interactives. Case in point: in “How Different Groups Spend Their Day,” The Times visualized a survey that asked thousands of Americans to recall every minute of every day.27 The interactive looks a lot like many of the interactive data visualizations The Times does today, only now they have to rely on actual programming to create them. This moment—the demise of Flash—helped further solidify the importance of designating a specialized set of people in the newsroom who could create the now more widely used and increasingly sophisticated interactives that had become more and more a part of the user experience online. Ajax But most of what we think of as being essential to the experience of inter- actives—the ability to move an object on a Web page without refreshing the page—is thanks to Ajax, a combination of technologies that eliminated the need for a “start-stop-start-stop nature of interaction on the Web,” as user-experience guru Jesse James Garrett explained in a 2005 essay that popularized the name Ajax. He explained the innovation further: “Instead of loading a webpage, at the start of the session, the browser loads an Ajax engine—written in JavaScript and usually tucked away in a hidden frame.... The Ajax engine allows the user’s interaction with the application to happen asynchronously—independent of communication with the server. So the user is never staring at a blank browser window and an hourglass icon, waiting around for the server to do something.”28 Ajax wasn’t just about visual images (you can see what it does by comparing your standard Gmail to the html view). But it meant that a program like Google Maps could enable clicking and dragging of the map without a user having to wait around for a page to reload. This single improvement in the design of interactives offered not only a better user experience but also made them truly interactive, because the interactive gave a seemingly immediate response, truly browsable and searchable without complications, and generally put the experience of the interactive in the hands of the user in a way that helped to tell stories in a seamless fashion. This advance helped interactives really get the attention of the news industry in 2005. As DeVigal remembers, “Ajax made things on the page dynamic. You could suddenly start typing in the same space and it The Rise of a Subspecialty 51 would work... it was mind blowing. Next thing you know, everyone started doing Ajax.” But the advances of Ajax represented a double-edged sword; now, the new realities of Web design and Web programming had suddenly made it more difficult. Java was hard to build and didn’t necessarily work properly on every platform. Web design had once been about graphics and layout; it was now about programming and, as Ethan Zuckerman put it, was responsible for “turning the web design job into a highly technical, programming job in some cases.”29 Social Open Source The open-source software movement has been around since the 1970s and early 1980s but perhaps got its true start when Richard Stallman founded the Free Software movement in 1984. The ethos of this movement was that if code was open, people had freedom to use this code, modify it, and share it—either after making changes or in the code’s original form. The underlying spirit was that “coding in the open”30 would encourage community, provide transparency, and let others build new and better projects on the structures of existing projects, as well as refine and fix projects, making software develop- ment faster and better. But for journalists working in newsrooms, the rise of open source didn’t really make much difference until the early 2000s, when platforms were developed to host open-source code in a way that helped promote social networking. Some of the benefits of open source for the newsroom (and elsewhere) were explained to me as critical to the widespread growth of interactives. Nathan Ashby-Kuhlman, who began his career at small Florida newspapers and now works at The New York Times, explained: Smart people were having to solve these problems over and over again. In the second half of the 2000s, social open source only started to emerge. But we were starting to do the pattern recognition and say, I see a repeatable way to save myself a heck of a lot of work, so instead of the repeated task of building a framework, I can use open source. It was a prerequisite to doing more of that type of work [interactives] at scale.31 Adrian Holovaty echoed this point—and actually called open source the “biggest thing” to influence interactives in journalism: “It’s free, really high- quality software that does all the bookkeeping and boring stuff for you as a programmer, so you can focus on building your site, and because of that it became much much faster to develop high-quality [work].32 Sharing code 52 Chapter 2 enables faster development time, exposure to new ideas, and a shared com- munity of interested parties seeking to improve news software (or app) de- velopment more generally. Open-source repositories had been around for decades (much of what we know about the Web is built on open-source software: Linux, Apache, Perl, MySQL), but the social open-source com- munity may have truly blossomed in the first years of this century, a time when newsrooms discovered these open-source repositories and began using them for development. Open-source repositories that allowed for a social dimension were critical, as Ashby-Kuhlman notes, for giving interactives the capacity to scale across newsrooms big and small. Depending on whom you ask, the social open-source experience can be traced to a number of starting points. Brian Boyer of NPR argued that it began first with Source Forge in 2000 (though the principles of social open-source platforms go back decades earlier to these early open-source repositories).33 Source Forge was the first platform that was a truly large repository (or place to host code) that allowed people to see what other people were doing in an easily navigable way and in a variety of programming languages. Google Code, in 2006, was also widely adapted by many in the news industry. And by 2007 it was clear that open source had become integrated into the social networking movement more generally. Mashable, for instance, wrote an ar- ticle titled “Open Source Social Platforms: 10 of the Best.”34 Now, most newsrooms use and collaborate on GitHub, an open-source repository that began in 2008 and came into its own a few years later. Almost every news organization discussed here has a profile on GitHub that hosts at least some open-source code. As Boyer noted, “Now, you’re standing on the shoulders of giants,” referring to the fact that the best minds creating software are sharing their work and making it possible for others to create high-quality work with similar utilities—and perhaps improve on them. Cloud Computing Interactives take up space on servers, demand computing power, and may not work well with the rest of a news organization’s technical infrastructure. And getting space on physical newsroom servers used to require careful negotia- tions with higher-ups that could slow down the development process. The rise of cloud computing reduced the price of Web servers, gave interactive journalists more computing power to work with, and helped them bypass the organizational and technical problems of their organizations. As Holovaty explained, “If you wanted to run a site before 2005, you had to pay for a Web server. You had to physically buy a computer and put it into The Rise of a Subspecialty 53 a server rack and pay for it. Now with cloud computing you can rent out a computer by the minute or hour, and if you get a lot of traffic, you can spin it off to extra computers on the fly. [Cloud computing] solved a lot of the technical challenges of the infrastructure.”35 So cloud computing essentially made the software-development process more flexible and more powerful. Ashby-Kuhlman added that before cloud computing, “you could either try to become best buddies with people who maintain the webservers or take over the system, which was maintenance, not journalism.”36 As such, cloud computing was just another technology that helped bring interactives to the forefront of the newsroom, as interactives could be more powerful and plentiful. Experiments could take place without having to garner hierarchical newsroom support, ultimately making interactives better. Overall, interactive journalists were able to marshal the benefits of new technologies to support their work. Each technology helped underscore the differences in interactive development from the rest of the newsroom pro- cess—yet while these technologies helped ultimately make the news website better, it was interactive journalists who were directly making use of them to change the way that news itself looked. Using and understanding these changes and their potential separated interactive journalists as having a spe- cial claim on knowledge and a specific utility in the newsroom from other professional journalists who did not. A Turning Point: Adrian Holovaty and Chicago Crime Maps Holovaty, who was working in Chicago for The Lawrence Journal-World, cre- ated a project called chicagocrime.org in 2005 in his spare time. The project was called the first news “mashup,” as it combined two sets of data—Google Maps plus data published by the Chicago Police Department—and made it interactive. The effort cemented Holovaty’s reputation and resulted in loads of press attention; consequently, he often gets recognition for being the first true hybrid—as both a programmer and a journalist. In a blog post announcing the project, he described the site: “[It] is a freely browsable database of crimes reported in Chicago. My scripts collect data from the Chicago Police Department once every weekday. The site slices and dices crime information in a ton of different ways, complete with a wide assortment of Google Maps.”37 It takes some imagination to remember what life was like before Google Maps was widely used, but think of a world that had not yet seen what most people consider to be a truly basic function: think about what Yelp can do—how it has restaurants plotted on a Google Map, for example. 54 Chapter 2 Holovaty—first with his project that put Chicago Transit maps on Google Maps, and now with this journalism effort—was the first person to show it was possible to see data plotted on top of maps through a truly dynamic experience. He explained in the post what some of the site’s unique features were: The map view lets you view crimes by a number of criteria, all updated dy- namically on a Google Map via Ajax. Find your district uses the Google Maps interface to guess which police beat you’re centered on. Every city block in Chicago has a detail page with its latest crimes and links to crimes within 1, 2, 3, 5 or 8 blocks. It’s got RSS feeds for every block and police beat in the city.38 This was the most technically sophisticated yet easy to use interactive that the news industry had yet seen. Holovaty garnered some serious recognition from the industry. He re- ceived the 2005 Batten Award for innovation in journalism (now the Knight- Batten award). He was interviewed by Editor and Publisher, the American Journalism Review, The Online Journalism Review, Spain’s El Pais, and was covered in The Chicago Tribune and The New York Times. The New York Times praised chicagocrime.org in its 2005 Ideas section: The most influential mashup this year wasn’t a Beatles tune remixed with hip- hop lyrics. It was an online street map of Chicago overlaid with crime statistics. ChicagoCrime.org, which was created by the journalist Adrian Holovaty, was one of the first Web sites to combine publicly available data from one site (in this case, the Chicago Police Department’s online database) with a digital map supplied by another site (in this case, Google).39 This was big praise from The Times, which credited Holovaty for setting in motion the mapping of real estate, classified ads, sporting events, and movie and gas station pricing. The Chicago Tribune caught up with Holovaty in 2008 after he turned chicagocrime.org into a startup called EveryBlock (which was later bought by NBC). The praise was effusive: Holovaty [is] part programmer, part journalist.... [He] and his team have fash- ioned a site hailed in both tech and journalism circles for its multiple innovations. Although it’s still fighting, like almost every other new Web destination, to find a mass audience—and is even further from finding revenue—EveryBlock, many believe, will help define the future of journalism.... Holovaty is a sought-after speaker on the Web 2.0 conference circuit, where the people remaking the Internet gather to discuss the hows and whys of the The Rise of a Subspecialty 55  venture. There are philosophical and personal disputes within and between the many Web communities, but most everyone seems to agree on Holovaty’s merit and fundamental good-guy-ness.40 Holovaty was the prototype for the programmer who could also be a jour- nalist, someone who forms the backbone of interactive news teams today. An article by Mark Glaser in MediaShift in 2007 was perhaps the first to emphasize the importance of hiring programmers to work on editorial teams. Whenever journalist-programmer extraordinaire Adrian Holovaty speaks at a conference, newspaper executives approach him to ask, “Where can we find another person like you?” Unfortunately, not a lot of people combine journalism with computer programming to create mash-ups like Holovaty’s seminal side project, ChicagoCrime.org, which feeds the city’s crime blotter into a searchable online database and onto Google Maps. Holovaty has repeatedly called on newspaper editors to hire programmers, and many of them are finally heeding his advice and considering ways of getting computer programmers onto their news staff and out of the trenches of tech support or doing work on web classifieds.41 This article profiles how even small newspapers like The Tacoma News- Tribune and The Greensboro News and Record, as well as large newspapers like The Wall Street Journal, have hired what they call a “news programmer.” Though Holovaty had majored in both computer science and journalism, the focus of this article was on finding programmers who could come in and work in newsrooms. After all, as Glaser wrote, “As for teaching journalism students how to do computer programming, that’s a long way from happen- ing.” Today we see computer programming offered in some form—if just as an introduction to how to build a Web page—in almost every journalism school in the country. Glaser pointed out the cultural clash between programmers and journal- ists: pay and lack of experience in journalism versus communicating an editorial vision. And some editors worried that their teams would be disap- pointed that they were wasting resources on programming rather than hiring reporters, though journalists quickly came round. What was notable, though, was Glaser’s insistence that programmers would come work in newsrooms even though the pay was poor. He argued that the job would appeal to them because of the creative freedom it would allow, an undercurrent that occupies many contemporary hacker journalists’ motivations for leaving programming jobs behind. Holovaty explained on his blog what being a developer on a newsroom team could mean: “Sure, the money isn’t as good as a straight-up tech job, and the geek cred is nonexistent. But it’s worth it for the chance to 56 Chapter 2 be creative and to make a difference in your community. If you’re a hacker, would you rather be a cog in the machine or an independent voice?” This was perhaps the first time the idea of a hacker working in a news- room was expressed: the hacker programmer coming into the newsroom to be creative and serve a greater good, a reality that is now a critical part of interactive journalism. Hackers possess different ways of thinking about journalism and doing work (such as experimenting and iterating), and in chapter 3 we’ll see how hacker journalists articulate these ideas. But the un- derlying message from this MediaShift article highlights how programming was becoming something seen as essential to editorial work—and how a new trend in newsrooms was upon us. Holovaty was a defining figure for his innovations but also because he became a legend—not necessarily because he deserved it, but because the news industry was fascinated by what he represented. He was the first high- profile programmer journalist, the first person who really got programming and journalism and treated them both as equally valuable. As Rob Curley describes, “Adrian Holovaty was a big deal. This was first journalist I ever met who could write sentences and write code, and that made him really power- ful.” The coverage Holovaty received and his constant presence on the news- innovation conference circuit helped spread the message that programming and journalism were complimentary, necessary, and critical to the editorial project. As such, he set the example for news organizations about the type of people they wanted to work in their newsrooms—a critical development that essentially helped spur the evolution of the subspecialty across the industry. Rich Gordon and the Creation of the Knight Foundation Scholarship In the mid-2000s, the prevailing attitude seemed to be that programmers had to come into the newsroom and help with interactives, and the widespread assumption was that most journalists simply didn’t have the skills to do the complicated programming required to produce interactives. Even though newsrooms were working in Flash, a tool that required less know-how with fairly admirable results, and there were staffers who were programming and had been trained as journalists, the consensus seemed to be that programmers would help bring the best of technological innovation into the newsroom. With this in mind, Rich Gordon, a professor at Northwestern University, attempted to find a way to create an army of Adrian Holovatys: journalists who could program. In his 2006 initial application to the Knight Foundation’s News Challenge Grant, he wrote: The Rise of a Subspecialty 57  Adrian Holovaty, now of The Washington Post, is widely recognized for the innova- tive projects (Lawrence.com, chicagocrime.org and Washington Post projects on congressional votes, political campaign ads, etc.) that combine his expertise in computer programming with his journalistic understanding and commitment. His understanding of technology, melded with a journalist’s sensibility about what’s important to people and our society, gives him the ability to find stories in data and to recognize opportunities to make data valuable to media consumers. The journalism industry needs more journalist/programmers, but there aren’t many to be found. [Northwestern University’s] Medill [School of Journalism] is well positioned to start turning them out.42 Gordon had been invested in the combination of computing and journal- ism since the late 1980s, when he attempted to use computer-assisted report- ing techniques on a Lotus 123 spreadsheet using the one computer in the newsroom. When he learned about the Internet in the early 1990s, he began thinking about how programming might help newsrooms. But at the time, he was mocked: when he was working at The Miami Herald in the mid-1990s, a higher-up at Knight Ridder castigated him for hiring a developer. Gordon recalled, “He said, you don’t need programmers, you just need journalism.” As a long-time advocate of developers in the newsroom, Gordon used Holovaty’s stature and influence to garner enough support within the news industry to persuade newsrooms to bring in programmers. But Gordon also saw the value of giving these programmers journalism training in order to avoid cultural conflicts and to make them not only able to understand the edi- torial process but also able to think editorially. Northwestern’s M.A. in jour- nalism had been fashioned and marketed as a one-year program that could train anyone, without any journalism experience, from fashion designers to artists to scientists, how to be a journalist—and would train programmers how to become journalists, too. As Gordon wrote in the Knight application, “Every year our one-year master’s program takes dozens of students with little or no journalism experience and turns them out ready to start working in newsrooms.... It is an ideal program for someone, such as a computer science major, who did not study journalism as an undergraduate.” Gordon’s Knight News Challenge proposal called for money to fund nine full scholarships (three awards over three years) with a $25,000 stipend to journalists who have undergraduate degrees in computer science, with a goal to then place them in jobs and internships in the news industry. He proposed that the buzz from the program could have a wide-ranging effect on encouraging other programmers to think about journalism as a viable career option. He wrote, “Publicity about the program will also lead other young technologists to consider journalism as a possible outlet for their 58 Chapter 2 talents, leading to more applications—to Medill and other schools—from skilled programmers seeking a journalism degree.” Gordon was awarded $639,000 for the scholarships. These nine scholarships were eventually funded with $900,000. The win- ners have gone on to have an outsized role in interactive journalism: the first winners included Brian Boyer and Ryan Mark. Boyer coined the term “hacker journalist” to describe programmers who could also do journalism, and Mark headed the Chicago News Applications team (after Boyer) before leaving for Vox. Others have gone on to work at The Palm Beach Post, The Washington Post, and in public radio; one co-founded Narrative Science, one of the first algorithmically generated news-content creators. The program has been successful in attracting and producing hacker journalists, though, as we will see, journalists have also taught themselves how to code, and they now form the bulk of interactive journalism teams. The scholarships helped further define this growing subspecialty. The cre- ation of institutional funding and training suggested that there was a defin- able type of person who could become an interactive journalist, with a clear background and skillset. Similarly, professions have traditionally been defined in part by whether specialized education was necessary for employment in the field, and in this case, the programmers received a specialized education about journalism. The idea of a specialized education in helping to create the profession is more convincing when reversed, when journalists in journal- ism schools enroll in computer science programs or enroll in a blended CS/ journalism program. Columbia launched an M.S. in computer science and journalism in 2011, intended for programmers with previous experience; in 2014, Columbia introduced the Lede program, a twelve-week or twenty-four- week nondegree program intended to give journalists “computational skills needed to turn data into narrative.”43 The program announcement advertised: Data, code and algorithms are becoming central to research and creative work, and are setting new parameters for the exercise of responsible citizenship. Co- lumbia’s Graduate School of Journalism and Department of Computer Science have together created two new post-bac certification programs that will offer hands-on training in data and data technologies, all taught in the context of journalism, the humanities and the social sciences. These programs assume no prior experience in these topics and, in fact, are explicitly aimed at students with little or no formal training in computation and data.44 This program began enrolling journalists who had no programming skills in basic computer science classes. Medill began training programmers in 2008, but since 2012 has been facilitating journalism students’ wish to double major The Rise of a Subspecialty 59 in computer science. Similar efforts have emerged throughout the country as journalism students either receive training in how to code as part of their core journalism curriculum or find that majoring in computer science and journalism has become easier, and more desirable. Though this sort of train- ing does not serve as a specific entry requirement into the field, as some might think necessary to define a profession or subfield, it does signify the kind of specialization necessary and the way education helps to specialize both programmers and journalists. In many ways, Gordon popularized the idea that a journalist could learn to code and a programmer could learn journalism within an educational setting, underscoring just how distinct this subfield was from other forms of journalism. Key Moments at Big Papers As interactive journalism came into its own from the mid-2000s through today, there were a couple of key moments that helped convince leading newsrooms of the potential of these projects and the importance of having people who can program on staff. Smaller newsrooms have employed pro- grammers in one fashion or another for as long as the big ones, but the efforts of big newspapers were what got the attention of the industry. Moreover, big newspapers, for better or worse, codify and reinforce existing trends: even if they do not set them, they compound their influence. The New York Times and The Washington Post are the principal newsrooms where interactive successes have helped establish the potential of interactives, not only in their own newsrooms but also to the industry at large. Each newsroom had key moments in the 2000s that set into motion the support for the large teams that exist in their newsrooms today and further signaled the importance of interactive journalists to newsrooms at large. The Washington Post had the early initiative in terms of interactive journal- ism. Holovaty came to The Post in 2005 and started a separate site designation (now defunct) called Post Remix. The site hosted the initiatives of outside developers who used Post data to create projects.45 Holovaty was also respon- sible for pioneering political infographics that same year, when he and his team (including future prize-winning New York Times interactive journalist Derek Willis) created “Congress Votes,” a database that let users browse every vote in Congress since 1991 in a number of ways, as Holovaty explained, “such as votes that happened after midnight, vote missers and, on a lighter note, vote totals by astrological sign.”46 Amusingly, the blog post Holovaty wrote announcing the project included a direct shout-out to a search for “the page for Barack Obama.” 60 Chapter 2  And in 2006, Holovaty would create “Faces of the Fallen,” an interactive database that would allow users to search through a browsable database of U.S. service members who had died in Iraq and Afghanistan. He notes that the first version of the interactive was in Flash, but that he had used Django (the Python-language framework he created) to improve its look and func- tionality. Holovaty described the project: “The site lets you browse by age, death date, home state and city, military branch or multiple search criteria. Each soldier gets his or her own page, as does each date, American city, age, military branch, etc. There’s an RSS feed for recent casualties, a feed for each state and a feed for each military branch. We’ve integrated Google Maps on several pages to highlight service members’ hometowns.”47 The project drew attention from industry publications and mainstream press. And to provide further evidence of its effect on the industry, The New York Times soon fol- lowed with its own version of chronicling war casualties, “Faces of the Dead.” These projects at The Post that got attention (and replication) positioned the newspaper as a forerunner in interactives. With Holovaty, The Post was in many ways setting a precedent for the rest of the industry. An interactive team in the newsroom could produce compelling projects that were alternative yet effective ways to tell stories. The projects enhanced and created news value, and captured the attention of the news industry. For the first time, on Web sites of a big news organization working in the age of broadband, users were in control of the information they could access once they arrived at a specific piece of content. Storytelling was changing, and so were the storytellers. At The New York Times, a field trip and the 2007 collapse of the I-35 bridge in Minneapolis set in motion what would become The New York Times’ In- teractive News Technologies Team. Head of the team, Aron Pilhofer, went on a trip with other key digital staff to show and tell Google about The Times’ digital products and to see the latest from Google. The Web giant was dismissive of journalism’s efforts, as even The New York Times didn’t have much going in terms of cutting-edge digital endeavors at the time. This insult made The Times start thinking about how it could be more ef- fective in a digital format. And when the Minneapolis bridge collapsed in 2007, Pilhofer and Jacob Harris, a programmer who had started working at The Times after leaving a financial-services firm, began thinking about how they could marshal the data about the bridge conditions. But, as Harris remembers, “We wanted to have everything up with what conditions and we couldn’t do it in the time we had.”48 For The Times and most other news organizations, though, presidential elections require advance preparation. And creating at least some interac- tives for the presidential elections could be done in advance. The small team The Rise of a Subspecialty 61 Pilhofer had assembled created projects for the primary and general elec- tions with campaign information. But the most significant effort was the team’s response to the release of Hillary Clinton’s White House calendar. Harris remembered, “Her campaign claimed that she did all these things,” but whether she had or hadn’t would require going through the calendar documents. The Times digitized the documents and created an interactive PDF viewer (a project that would later become Document Cloud, a collab- orative document-sharing platform). Now, The Times’ reporters and editors and ordinary readers, too, could sort through Clinton’s calendar and help spot inaccuracies. This project showed The Times how interactives could contribute to breaking-news projects and helped justify devoting resources to interactive efforts. The 2008 election prompted the creation of other key tools, including one called Puffy, which enabled users to submit photos of the presidential inauguration and which The Times could moderate and then post. The effect was that inauguration coverage was augmented by hundreds of user photos capturing photo angles, crowd shots, and emotions that The Times’ cover- age team could not. Interactive journalists could also build tools, thanks to their programming skills, ones that clearly had the capacity to enhance news coverage. The case for the interactive team, however—for bringing on people with special skills and knowledge and for giving them the opportunity to contribute to big news events—was brought into sharper focus after the 2008 election. The Times continued to work on interactives, but it wasn’t until 2010 that it became clear that interactives were a critical part of The Times’ claim to prestige and recognition. That year, Matt Richtel, a tech journalist, had been working on a series of stories about distracted driving, or what happens when people use their cellphones and other electronic devices behind the wheel. Key members of The Times’ staff with programming experience built an interactive game: while navigating a computerized version of a cellphone, you were also meant to use your cursor to navigate a video-game-style street made in the image of the Nintendo game Paper Boy, with obstacles and all. It was essentially a video game built as part of the Pulitzer package, but it drove home (literally), in a way that a story alone could not, that distracted driving could get you killed. Richtel won the Pulitzer Prize that year for national reporting. The Pulitzer committee gave a nod to these efforts, granting Richtel and The Times the award for “incisive work, in print and online, on the hazardous use of cell phones, computers and other devices while operating cars and trucks, stimu- lating widespread efforts to curb distracted driving.”49 And when Richtel gave 62 Chapter 2 his thank-you speech to The Times’ staff, he acknowledged the interactive team, saying, “And, we have video game making skills. Old-world journalism is the essence of new-world journalism. The series was long form [journal- ism] with video, audio, and, yes, video games.”50 The interactive team had now been, in part, responsible for a Pulitzer Prize. The interactive news department had truly arrived: for better or worse, The Times is in a quest to win Pulitzers, and the interactive team was now recognized as a critical part of this mission. Elsewhere around the news industry, it was also clear that The Times’ Pulitzers, which in 2010 included investigative and national reporting awards, also commended interactive and multimedia elements recognized by the Pulitzer committee as critical to the success of the report- ing efforts. This kind of external validation—these prizes—are critical for securing a sense of occupational identity, and in this case, the recognition reinforced the growing sense of both importance and self-definition of in- teractive journalists at The Times. And perhaps, in the age where traffic ultimately decides how much ad- vertisers are willing to spend on online news sites, and as news sites are struggling to survive, one linchpin for the subfield’s arrival in newsrooms was evidence between 2012 and 2014 that interactives were traffic drivers. At The Times, in 2012, there was “Snow Fall” to show the potential of bringing new visitors to The Times. As mentioned in the introduction, this included 2.9 million visits for more than 3.5 million page views (each visitor was read- ing more than one page). And at any one point during the story’s peak, of the twenty-two thousand people looking at “Snow Fall,” about seventy-five hundred of them were new visitors to The Times. That kind of ability to attract new visitors is critical to the site’s growth. Subsequently, in 2013 and 2014, interactives were major drivers of traffic over the course of the year on The Times site—in fact, The Times produces so many interactives that members of the interactives team, graphics team, multimedia team, and others might be responsible for them (“Snow Fall” was actually technically a project of the graphics team, not the interactives news team). A quiz that tried to identify your dialect garnered the first spot on the 2013 list—it was the most viewed story of the year—even though it wasn’t a story per se, but an interactive that helped you learn about regional dialect variation.51 In 2014 eight interactives made the top-twenty list of most-visited stories.52 The most-visited story of 2014 on nytimes.com and the mobile site was a photography-focused interactive slideshow featuring two sisters who had been photographed forty times over the course of forty years (“Forty Portraits in Forty Years.”) The Dialect Quiz came in again at third. Then in fourth, “52 Places to go in 2014”; in eighth, “Ebola Virus Q&A” (complete with a map of The Rise of a Subspecialty 63 “52 Places to Go in 2014,” The New York Times Ebola cases in the United States and in Africa, and some reassuring facts); in tenth, “Where are the Hardest Places to Live in the U.S.?” (Clay County, Kentucky, was number 1); in eleventh, a breakdown/explainer of a popular video game (“10,000 League of Legends Games in 30 seconds”); in sixteenth, “Is It Better to Rent or Buy?”; and in eighteenth, “The Ukraine Crisis in Maps.” Other top traffic drivers on other measures (such as most-shared on social media) included “Mapping Migration in the U.S.” (which tracked where people were born versus where they moved) and “The Premiere League If 64 Chapter 2 Table 2.1. “The Year’s Most Visited Articles, Blogs, Multimedia and Interactives” 2014 Rank Page Title Interactive? 1 Forty Portraits in Forty Years Yes 2 An Open Letter from Dylan Farrow 3 “How Y’all, Youse and You Guys Talk” Yes 4 The 52 Places to Go in 2014 Yes 5 “Philip Seymour Hoffman, Actor of Depth, Dies at 46” 6 What You Learn in Your 40s 7 For the Love of Money 8 Ebola Virus Outbreak Q&A Yes 9 “Robin Williams, Oscar-Winning Comedian, Dies at 63” 10 Where Are the Hardest Places to Live in the US? Yes 11 “Watch 10,000 Leagues of Legends Games in 30 Seconds” Yes 12 Autopsy Shows Michael Brown Was Struck at Least 6 Times 13 The Scientific 7-Minute Workout 14 Thanksgiving Recipes across the United States Yes 15 Jaden and Willow Smith Exclusive Joint Interview 16 Is It Better to Rent or Buy? Yes 17 Suicide Bomb Trainer in Iraq Accidentally Blows Up His Class 18 Ukraine Crisis in Maps Yes 19 Doctor in New York City Is Sick with Ebola 20 The Horror Before the Beheadings Source: nytimes.com and m.nytimes.com, adapted from nytimes.com data. Only English Players Counted” (an entertaining portrait of globalization in soccer). Out of twenty stories that referred traffic in a single year on the Web site and mobile site, eight were interactives; this means that almost half of all the top traffic referrers to The Times were interactives in 2014. You can see a mix of entertainment-focused interactives, data-focused interactives, and some directly news-related interactives. This suggests the wide-ranging appeal of interactives. This kind of external measurement of success underscores just how important interactives were becoming to The New York Times’ overall economic health and its future plans. Interactive journalists had the case for their work’s importance made through the popu- larity of interactives. The reification of interactive journalists as not only a specialized subfield but also as a respected and critical part of the newsroom had truly come to fruition at The Times, and this effect was also being seen throughout the industry. A few news sites have seen tremendous success with news quizzes and games. These are often dismissed as not-serious news, but they may be related to a news event, or they may provide useful information, or at the very least, they are bringing new people to the news site who may never have been there before. Slate, for instance, received its most traffic ever for a quiz that mocked The Rise of a Subspecialty 65 John Travolta’s attempt to say the name of a Broadway star at the Academy Awards (called “Travoltify Your Name”). People had fun, they shared it, and they came to Slate. Time magazine’s 2014 quiz about how much time people waste on Facebook got 4.7 million hits and gave Time its biggest traffic day ever, with 3.8 million uniques. Yes, this was not a strictly “news” interactive, as some detractors might suggest, but it contributed useful information to people, offering guidance to them about their world.53 Quizzes aside at Time, a Digiday story notes that a new leader of the “inter- active data team” was brought on (at Time) in 2013, and “the stories his team have produced have been Time.com’s most popular for three years running.”54 As a result, Time has even created a separate page to house all its interac- tives, called “Time Labs,” in order to maximize the visual exposure that the interactives can get, though for maximum impact the interactives will still appear on Time’s site. The justification for the new site was put in financial terms, according to Digiday: “Advertisers are gaining interest in how much time readers are spending on a site versus just clicks, and Time Labs also is a way for the publisher to capitalize on that interest. Buyers said more time spent can lead to a greater advertiser benefit, which in turn could help a site command higher ad rates, or at least more advertising.”55 As such, interactives are critical to newsroom survival—tied to economic success and seen as a way to marshal audience attention, throughout the news industry. And the people who build them have begun to matter more and more to the success of these news organizations. The importance of interactive journalists as a significant subfield within journalism has been enshrined by these external measures of success they now achieve and the critical role they may play in the survival of these news organizations. The Beginnings of Data Journalism through Computer-Assisted Reporting In this case, it’s also helpful to offer a quick history of how data journalism came into its own in today’s newsroom. We can see through this historical lens how data journalism is a bit different from the overall project of interac- tive journalism. In the eighteenth century, tables in early American news- papers kept track of stock prices and the import and export of commodities into ports.56 The first issue of The Guardian newspaper featured data charts on public education. More complicated tables emerged in the early, special- ized business journals, particularly with the rise of The Wall Street Journal in 1889. Early box scores for sports appeared at the end of 1870s. And by 1896, maps with electoral information appeared on front pages. 66 Chapter 2  C. W. Anderson argues also that some form of data journalism has been present in journalism throughout its modern history. He notes a variety of instantiations of data in journalism: from the turn away from documents to oral reports in the age of the penny press, to the turn to social science in the 1960s, to the early 2000s and the interest in building patterns through mathematical models. He argues that the survey movement of the 1900s offered new techniques for visualizing and collecting social data, which in turn inspired journalism. 57 Computational journalism, defined broadly as the application of com- puter science methods to journalism (or more specifically, using algorithms, data, and knowledge from social sciences and applying them to journalism through computing technologies)58 only emerged in the 1950s. Melisma Cox argues that the first instance of computer-assisted reporting (CAR) in the United States was in 1952, when CBS used the Remington Rand UNIVAC to predict the outcome of the U.S. presidential race between Eisenhower and Stevenson.59 Matthew Reavy also points to innovators such as Philip Meyer, who in 1967 used an IBM 360 mainframe to analyze survey data about the Detroit riots.60 Other early-adopter journalists began using computers in their work to analyze data. In 1973, Meyer wrote Precision Journalism, advocating for the greater integration of computers, data, and social science methods in journalism practice.61 That same year, in perhaps an early iteration of open-source ethics in journalism, The New York Times made public an interactive information system with data about New York City police statistics.62 Around this time, famed Philadelphia Inquirer reporters Don Bartlett and James Steele, with help from Meyer, put court records into a computer for their series “Unequal Justice.”63 Possibly the most visible moment for computational journalism came in 1989 with Atlanta Journal Constitution reporter Bill Dedman’s Pu- litzer Prize–winning report on unequal housing practices and red-lining. An anecdote from the late 1980s underscores the complexity of retrieving information from databases. Elliott Jaspin, a Providence Journal reporter, just wanted to do one thing: get government data and use it for stories. But the information hadn’t come in paper form, unfortunately; no, in 1987, the U.S. government was storing data on nine-track tapes resembling film reels that could be read only by giant mainframe computers (a common practice for large organizations at the time). Jaspin had access to his newsroom’s main- frame computer and had some success using state databases, even finding information that led to the head of a state housing agency going to jail. But many smaller newspapers did not have access to mainframes—and journal- ists who could access mainframes at larger newspapers were often uncom- The Rise of a Subspecialty 67 fortable using Unix and EBCDIC (used in binary file encoding). He wanted to make accessible to other journalists what was then a modern method for using databases. Thanks to a fellowship through Gannett at Columbia University, Jaspin learned enough about nine-track tapes that he could access the data through a PC. Jaspin said he thought, “If I could write the software that would allow a reporter to read and download the data to a PC, I could mimic a mainframe for $10,000.”64 With Dan Woods, one of the journalism students who had a background in computer science, he wrote the program in nine months. But it was still too complicated for most, and special training was needed. Jaspin explained that there was still a “serious learning curve. How do you use relational databases? How do you convert files from EBCDIC to ASCII? What are variable length records and how do you download them to a PC?” So Jaspin approached the University of Missouri, and thus began the dis- cussions about what would become NICAR. The group was a subset of the Investigative Reporters and Editors professional group and was founded in 1989 with the goal of helping journalists extract, analyze, and report on electronic information.65 By 1991 the second edition of Meyer’s book argued that journalists inter- ested in CAR were a separate breed of reporters who often bought their own computers in advance of newsroom technology. Journalism scholars began writing about the CAR movement around this time, and in 1996, Brant Hous- ton identified three key aspects of technological innovation and journalism production: database reporting, spreadsheets, and online-reporting.66 Other articles in the late 1990s and early 2000s function as scholarly discoveries of the increasing prevalence of the CAR journalist; at this point, though, computer-assisted reporting was still looking for stories through anecdotal pieces of data, rather than being able to use raw computing to find overarch- ing patterns to process large datasets. But CAR is only part of the data journalism equation. CAR speaks to the computational aspect, whereas the application of data journalism helps as- cribe additional importance to the presentation of data and how to examine data beyond anecdotes to a more comprehensive analysis of the entirety of the data. Chapter 3 will discuss exactly how data journalists understand and depart from CAR, and how these journalists are connected to but not always a part of interactive journalism. 68 Chapter 2 From the History to the People From Chicago and South Florida in the mid-1990s to Lawrence, Kansas, in the early part of this century, to The New York Times’ most-viewed list of 2014, the history of news interactives has developed to the point where interactive journalism and journalists are now a regular part of newsrooms. The subfield of interactive journalism has emerged thanks in part to changes in technol- ogy, to changes in economics, and to claims to the ability to do particular kinds of work. These journalists have also been accorded cultural power in the newsroom as their work has developed at scale, grown more sophisticated, and demonstrated success online and on mobile devices. As such, they have also been able to influence news organizations with their work. Unlike other subfields, there isn’t much contest between interactive journal- ists and the rest of the newsroom. Certainly, there may be difficulties when interactive journalists and journalists who don’t understand how software works try to collaborate. But with everyone in the newsroom wanting their story to be “snowfalled,” and with news organizations investing more resources in interactive departments and interactive journalists, it becomes only more obvious what practical and abstract claims to knowledge these journalists have and what they can offer to the newsroom. Thus, we turn to who the people are who actually make up the subfield of interactive journalism. The Rise of a Subspecialty 69