Skip to main content
simon houpt

A piece of software made Travis Solorio a little bit famous after he was gunned down on the streets of Los Angeles in February. On the morning following his death, the LAPD sent out an e-mail alert about the homicide to newsrooms across the city; within seconds, the Los Angeles Times posted a one-paragraph story about the murder. But no reporter or editor was involved: At the Times, when it comes to homicides, the first draft of history is written by a computer.

If your kid plays hockey or baseball, you may already be familiar with algorithm journalism. Last fall Pointstreak Sports Technologies, a Canadian company that helps sports leagues compile game statistics, struck a deal with Narrative Science, a Chicago-based tech startup, to enable stories to be automatically written on every single game in its database. Parents or players can post those stories – which read like slightly stilted versions of the conventional game reports carried by wire services – to Facebook, or send them to their community paper.

This year, Narrative Science figures it will generate about two million stories on youth sports. "Each of those stories may have an audience of 20 people," noted Stuart Frankel, the company's CEO. "But to those 20 people, it's the most important thing they're going to read that day."

All of which is freaking out some journalists, who are already in the midst of a well-chronicled existential crisis. But if my colleagues will take a moment and a deep breath, they'll realize the development is just another part of a painful process that is transforming journalism from a trade that traffics in information to one that helps in the development of knowledge.

Narrative Science, which grew out of research conducted at Northwestern University, opened for business in the fall of 2010; it already claims to serve about a dozen media companies, including Forbes. For one client, which it won't name, its software checks all of the publicly listed U.S. equities every few seconds: If a stock hits one of the trigger events the software is looking for – say, a new 52-week high – it will automatically generate a story.

"Literally within a couple of seconds there's a story that's generated, published, distributed via mobile application, and available for audience consumption," said Frankel. "We do hundreds of these stories a day. A reporter just couldn't do it. The data becomes stale very, very quickly."

This isn't the first time reporters have been faced with subjugation by software: More than 20 years ago, the financial information terminals leased by Bloomberg L.P. began offering robust number-crunching functions that were once performed by reporters and analysts. Once they got over the insult, reporters realized they could find stories buried in the data they couldn't see before.

And algorithms are already a fact of life in most modern news organizations. Just about all of the leading newspaper websites and portals use them to help determine the placement and rotation of stories.

Last month Ben Welsh, a database producer at the L.A. Times, told an audience at the International Symposium on Online Journalism about some of his paper's initiatives in computerized reporting, which include a program that scrapes LAPD arrest records for people in prominent positions such as ministers, musicians and actors. He noted that the same tools could be used to search for stories on money in politics, SEC filings, home sales, election results, environmental data and a host of other subjects.

"People often think there's a kind of binary relationship," said Frankel of Narrative Science. That is, there's a belief that stories "are either machine-written or human-written. But I think over time, what really becomes the norm is that a reporter is off researching his or her story, getting information that's not captured in some database somewhere, really understanding the story in total, and writing part of that story, but also including a portion of the story that in fact is written by a piece of technology.

"We free up journalists, frankly, to do stories that can't be done by technology – and probably will never be able to be done.

"Like this story."

Follow related authors and topics

Authors and topics you follow will be added to your personal news feed in Following.

Interact with The Globe