Making Data Journalism Sexy

What I want to talk about is training data journalists in these countries and in this professional context. I work with Internews, an international non-profit organization whose mission is to empower local media worldwide to give people the news and information they need, the ability to connect, and the means to make their voices heard. We’ve worked in over 90 countries, so we have the local media knowledge to tailor training programs that will prepare journalists to mine data, identify patterns, and investigate the reasons behind these trends, to cast a spotlight on development needs in their countries.

Our first attempt was based on our idea that the fastest and cheapest way to make data journalism—and, by extension, math—seem cool was through Data Journalism Boot Camps. A week-long training in the fundamentals of data journalism would give journalists just the push they needed to start fighting corruption through awesome visuals and news apps. But, just as a one-week conventional boot camp can hardly be expected to produce a special ops unit, the workshops did not produce a cadre of global journo-coders. In fact, many of them never produced a single data-driven story, despite a flurry of Twitter traffic that suggested that places like Nigeria, Nepal, and Bolivia are the next rising stars in data journalism.

Culture Matters

Part of the problem with our initial approach was that unlike in the US or Europe, our boot camp trainees didn’t usually have a support network after the training. They rarely had access to data journalism desks, chapters of Hacks/Hackers, or training institutions offering professional courses in data journalism. So when a trainee’s formula produced an unexpected error, the data was too dirty, or the visualization software didn’t support the local language, there wasn’t anyone to turn to for help.

Some groups have tried the opposite approach, emulating the OpenNews model and embedding lone coders in newsrooms. But without a data culture, the developers toiled away alone, and most data desks were manned by a lone intern. Journalists had never worked with data and the developers had never worked with journalists. The silos never magically disappeared.

Post-boot-camp survey data told us we needed to focus not on turning out data journalists, but on fostering a data community that could together overcome the many, many barriers to producing and publishing a single data story. So we decided to try to grow a data community first in the hopes that the community would then nurture the first generation of data journalists. (We also recognized that data journalists need access to data, and that was sorely lacking in the countries we were working in.)

Developing Data, Skills & Community

To grow our data community we needed both an online incubator for data access and offline incubator for training and community building. Our flagship site, the Data Dredger, is a resource for Kenyan journalists to download, embed, and publish visualizations of Kenyan data. Our first task was to accumulate enough data for journalists to tell stories relevant to their audiences. Many exercises in the training work relied on World Bank or UN data for country-level analysis, but this data is hardly compelling for your average Kenyan media consumer, who is more concerned with his or her family or community than with global trends; they want to know how many mothers die prematurely in their community as compared to one county—not one country—over. Journalists can access Kenya-specific visualizations on our site. We produced Politics of Health ahead of the 2013 Kenyan presidential elections to fact-check candidates’ campaign promises, and visualized areas of malarial risk, maternal mortality, and anti-retroviral shortages. The Data Dredger was the only finalist from Africa in the 2013 Data Journalism Awards.

Paul Wafula’s interactive for the Internews Fellows Blog.

By taking care of the tedious process of finding, scraping, cleaning, and verification—something that a journalist working on her own will not have time for—we created a shortcut for journalists, giving them a taste of open data. The visualizations started appearing in the Kenyan media, and journalists came to us for help with their data story ideas.

For our offline incubator, we decided on a four-month fellowship model where a group of talented media professionals would dedicate themselves to learning new skills and producing experimental content and storytelling for their home media outlets. We pitched the fellowship as a highly prestigious opportunity for journalists, graphic designers, and developers to develop specialized skills that would make their own news outlets shine. Luckily, the Kenyan media industry is huge and several outlets could afford to let go of a staff member for four months, which is not the case in a lot of developing country media markets.

For us, the fellowship served several purposes: it immersed the fellows in months of intensive training in which each module built up their cumulative skills, it introduced them to data that the journalists could use for real stories when they returned to their media outlets, and it encouraged them to work as a team to complete assignments. Most assignments were published on a fellowship blog.

Internews Fellow Michael Mosota’s blog post on data visualizaion

The fellowship blog encouraged fellows not only to produce but also to invest in creating an online brand for themselves. The blog built up their professional profile, disseminated Kenyan examples of data journalism through social media, and helped a data journalism community coalesce. It also enriched the Data Dredger by demonstrating the fellows’ evolving skills.

Establishing a brand and confidence in their work proved crucial when it came time for the fellows to return to their outlets with their data journalism projects. There, they had to fight not only for extra time to produce data stories but also for airtime, column space, and the ability to break out of their standard web templates to accommodate embedded interactive content and special project pages.

There was more to it than tech training, too. Halfway through the fellowship, we realized it was also about learning how to tell stories with data. Fellows went out to remote regions of Kenya to put faces to the data: find mothers who had lost daughters to complications of illegal abortions, men dying of preventable diseases when more county resources could have been spent on health, and people with disabilities who haven’t received safety-net payments in six months. The fellows submitted complete data analysis with findings about the demographics of those most in need. They also submitted traditional feature reporting that did not seek to find out why the system was failing, even though they had the evidence in hand.

We pushed trainees to ask hard questions to explain the data—not only to look for data to confirm their hypotheses, but also to see what conclusions they could come up with by focusing on the questions raised by the data. Where had the money gone for these programs, and what policy shifts caused uneven access to medical care?

The first version of an investigation by fellow Paul Wafula, a journalist for The Standard in Kenya, revealed a straightforward finding: most counties were budgeting less on health than they had been allocated the year before by the national government, even if their population was sicker and had fewer health facilities and personnel than average. After the training in how to integrate data and investigative journalism, Wafula uncovered the stall tactics, inexperience, and corruption at the county-level that led to under-budgeting and misspending, prompting a conversation about the skills required by the newly formed county-level governments in order to govern effectively.

New Digital Content for Traditional Media

Of course, the fellowship alone couldn’t transform a very traditional media apparatus and audience into data-lovers. For the first round of projects, the design of the print versions of each of the stories far outshone the online versions. In the Kenyan news cycle, page designers and television producers focus their attention and resources on traditional formats, and later an online editor uploads the content into a fixed template, often without the visuals, for a secondary audience.

Internews Fellow Mercy Juma fought to keep the infographics in her print story for the Nation.

As data floods the world, our fellows struggled to make it accessible to their audiences whose primary modes of media access are radio, television, and print newspapers. Kenyans are increasingly dissatisfied with the direction of their government and Kenya ranks in the bottom third (136 out of 175) on Transparency International’s Corruption Perception Index. But analysis and visual representation of government performance or development outcomes are practically non-existent in the media, which favors blow-by-blow reporting of political scandals. Many Kenyans have not attended secondary school and therefore many were never exposed to mathematical representations such as bar charts. Our end goal was a more informed citizenry, not to win awards for interactive visuals, so for the fellows’ final products, we went with easy-to-understand infographics and saved the bells and whistles for the web versions. Michael Mosota, our graphic design fellow, and Daniel Cheseret, our developer fellow, created two different versions: a version for print and an online version with more sophisticated news apps and interactive maps.

Institutionally, each of the fellows had to fight to defend their data journalism from being cut, transformed, or simplified. Mercy Juma, an Internews fellow and journalist for NTV, fought to keep infographics in her television piece and the print version that ran in the Nation (see above). She threatened to withhold her story if the producer did not broadcast the uncut version. In order to publish a news app that enables Kenyans to explore how much their county is spending on healthcare, our developer fellow Daniel Cheseret went into The Standard newsroom and set up an independent website that was later migrated to The Standard Online after the print story generated high traffic to the app.

Success: Fellows’ Stories Provoke Public Outrage and Government Action

Last month, the fellows saw some of the first concrete policy results of their investigations. The fellowship proved to several big media outlets that data journalism is worth exploring and sells in print, on television, and online.

“When the Sun Sets in Turkana; Hunger Stakes and Stripes in the North,” by Mercy Juma, ran as the lead news story on January 21. The 12-minute, data-driven story is longer than any lead story anyone can remember in the history of Kenyan television. Turkana is an isolated, impoverished region of Northern Kenya long neglected by the media and government. Juma’s story reveals that malnutrition in children is a growing problem in Kenya, as famine becomes more intense and frequent, and that money goes to emergency food aid, not long-term drought mitigation. Due to the massive reaction to the story from individuals and organizations—whose phone calls started flooding in before the piece had finished airing—the station established a relief fund for Turkana County within hours, as explained in the follow-up story: “Cases of Malnutrition on the Rise in Turkana Hospitals.” The fund had raised Sh1.2 million ($14,000 USD) by the end of January. Local politicians also offered to sponsor the families featured in the stories. The family she featured was also recruited into the safety net program that was the subject of investigation of fellow Paul Wafula’s story. The print version of the story “Famine Strikes Again” (PDF link) brought in more donations. She followed up the article with the TV story, “Hunger Keeps Children Away from School in Turkana.” Since then there have been prominent stories across the Kenyan media on the desperate famine situation.

Even more importantly for the fellows who worked on the story, the Drought Monitoring Committee asked Mercy to share footage from her story because they claimed they were not aware that the situation had become so desperate. They also requested access to her data, which she had obtained from another office also under the Ministry of Environment, Water, and Natural Resources after a long negotiation. The Committee explained that long-term drought relief strategies have been drafted but never implemented. Based on her water shortage data, the Ministry plans to travel to Turkana to dig more boreholes. The government, through the Ministry of Planning and Devolution, released 2.3 billion Sh ($27 million USD) to go towards relief distribution in Turkana County, a development that Mercy is following closely along with the progress of the longterm drought relief legislation. NTV is eager to send her on the expensive trek to the region for follow-up stories.

Keeping Up the Momentum

Data journalism is still in its infancy in Kenya, but it has moved forward. Our fellows have now established beats that give them more freedom to pursue data-driven stories about health spending, drought, poverty, and other key governance issues when they return to their media outlets. They presented their stories at the inaugural Online News Association Nairobi event, discussing their process with other digital journalists in Kenya.

Crucial to the success of the project was the gradual integration of the fellows into the global data journalism community. Civic hackers such as Juan Elosua from Spain—another country without access-to-information laws—provided crucial training and support in the absence of an active local Hacks/Hackers chapter or other places for journalists to get help from developers and join cross-border projects. Virtual spaces, like the School of Data, Source, and Data Driven Journalism can substitute for an in-country community, but only once journalists master the basics. Aurelia Moser, the 2014 Open News fellow co-hosted by Internews in Kenya and Ushahidi, is also helping integrate Kenyan journalists into the global data journalism community.

The Kenyan case is unique. Our incubator strategy worked because the project is well funded, the media industry is robust, and journalists are curious about digital media. Though there are currently no operating data desks at media outlets in Kenya, our fellows will soon grow into editorial positions where they can formally bring together journalists and the vibrant tech community in Kenya.

Data availability, data and digital literacy, and the media environment largely dictate the direction data projects can take in any given place. Over the next few months we will be launching data journalism projects focused on:

• the environment and food safety in China, where we will rely on an online discussion forum to provide support between weekend training sessions;
• a combination of a boot camp and mentored reporting projects for election reporting in Afghanistan; and
• partnerships between journalists and NGOs for human rights reporting in Nicaragua.

All these projects are focused on cultivating data communities that will prepare journalists to thrive, working within their local context with the resources they have available. It’s not within our reach to foster an entire new generation of journo-coders: what we can do is pool available talent from different sectors to coalesce into a data community. Working together with other local incubators, we hope to eventually offer fellowships in other countries to help journalists—like our fellows in Kenya—become analytical thinkers with a news-nose for data stories and a willingness to work with the nascent data community to get the story out, both in and outside the newsroom.