Text is a ‘new frontier’ in data journalism, says head of the IRE

If you want to be ahead of the curve in data journalism start thinking about how to make sense of the massive amounts of text out there. That is one of the “new frontiers” in data journalism, according to Mark Horvit, executive director of the Investigative Reporters and Editors (IRE).The organization held its annual computer assisted reporting conference this week in St. Louis (Missouri).

Mark Horvit

“One of the new frontiers in dealing with data is (…) being able to get a handle on how we can look at all the massive amounts of text that’s out there and make sense of that,” said Horvit during an interview at the St. Louis Union Station Marriott that hosted the conference.

Horvit believes that the challenge of dealing with text will get bigger because the web and social media make the amount of text to grow “exponentially.” That’s why this year’s computer assisted reporting conference included several panels with experts on text mining.

This year’s event also try to encourage the increasingly collaborative nature of data driven journalism, according to Horvit.

“We tried something that we have never tried before: to have a hackathon where people could use Scraperwiki – a non-profit organization that has created a tool for scraping data – to learn how to use the tool and then work on grabbing datasets while they are all together,” said Horvit.

There were also unplanned collective collaborative efforts like the one that took place on Thursday evening when people sat in circles with their computers at the hotel lobby. They worked together on techniques to take unstructured text and analyze it that were explained during the hands-on-session that morning.

“So this happened organically and it fits with a lot of what you see happening in communities around the country where people are just getting together and doing this sort of work,” explained Horvit. He thinks there is also the realization that “when you put multiple minds together you are more likely to come out with something better.”

The IRE decided to organize a hackathon after noticing that during last year’s conference participants were setting up impromptu sessions in the bar or the hotel lobby at night to hack on something or to work on a project or projects together.

During those evening sessions a bunch of data journalists started talking about the possibility of creating a tool that everybody could use to analyze the Census data. A few months later with the help of a grant from the Reynolds Journalism Institute eight journalists from competing news organizations developed Census.ire.org.

Looking to the future, Horvit says to be “very optimistic” about data journalism.

Collaborative spirit at IRE's conference in St. Louis

He explained that the ability to put data online, to make data interactive and the fact that often data brings traffic to websites are making people value more this field.

“I do believe there is a trend towards hiring more people with these skills, more desire in the newsrooms. You can look on our website. There are constantly job ads from newsrooms looking for people with data analysis skills,” Horvit explained.

Despite that, the head of the IRE insists there is still a long way to go: “We go around the country and do basic Excel training (…) and most journalists in every newsroom we go to for every regional workshop have zero experience using spreadsheets. Never mind anything fancier,” Horvit underlined.

The fact that a lot of people who get into journalism are math-phobic and the lack of knowledge about the tools that are available are “a big part of the problem,” according to IRE’s executive editor.

Add Comment Add yours ↓

Your Comment