Write a story with this data

Last week I hung out at Northwestern University’s Medill School of Journalism as I saw some of the groundbreaking collaborations its students are engaged in: primarily between computer science and journalism.

I get the sense that Medill is preparing for a reality of lean-and-mean newsrooms, with fewer paid staffers, allocating the grunt journalism to alternative sources of labor.  In this case, as we move deeper into the Data Age, where our actions, interactions and behavior are increasingly digitized into databases, what role will computers play in producing basic stories?  A pretty substantial one it seems: from easy-to-use applications to tease out interesting census stories, to online platforms that can fact-check congressional speeches.

And then there’s this one from Stats Monkey, which has just been spun-off into the startup Narrative Science (which already has as its first client, the Big Ten Network):

Imagine that you could push a button, and magically create a story about a baseball game.  That’s what the Stats Monkey system does.  Given information commonly available online about many games—the box score and the play-by-play—the system automatically generates the text of a story about that game that captures the overall dynamic of the game and highlights the key plays and key players.  The story includes an appropriate headline and a photo of the most important player in the game.

Does this render journalists obsolete?  It may relegate some journalistic tasks to computer processors, but in some ways, it makes the role of a human storyteller even more crucial: to build upon the story with further analysis, context, and drawing instinctive connections that artificial intelligence still can’t produce.

So what if you took the above box score, and clicked the yellow button “Write a story with this data?”  Here’s the story the computer would generate:

Joe Koshansky was strong at the plate on Saturday for Nashville.
Down to their last out, Koshansky saves Sounds

NASHVILLE — The Iowa Cubs had the Nashville Sounds down to their final out, but Joe Koshansky came through with a home run to drive home the winning run in Nashville’s 5-3 victory on Saturday at Herschel Greer Stadium.

After Josh Anderson singled and Ray Olmedo reached on a fielding error by second baseman Bobby Scales, the Sounds were trailing by one when Koshansky came to the plate against reliever Blake Parker with two outs and runners on first and second. He homered scoring Olmedo from first and Anderson from second, which gave the Sounds the lead for good.

Thomas Diamond was a workhorse on the mound for Iowa. He pitched six innings and allowed one earned run, five hits and two walks while striking out five.

Reliever David Riske got the victory in relief for Nashville. Riske pitched two innings and allowed two runs.

Nashville scored in two innings to claim the victory. The Sounds scored one run in the first and four runs in the ninth. In the bottom of the first, Diamond gave up one run on a single by Koshansky.

Parker suffered his second loss of the season for Iowa. He went two-thirds of an inning, gave up one walk, struck out one, and allowed four runs.

Iowa tried to spark a rally in the fifth and ninth innings. After Sam Fuld walked and Darwin Barney singled in the fifth, the Cubs were trailing by one when Chad Tracy came to the plate against Chris Waters with two outs and runners on first and second. He singled scoring Fuld from second.

In the bottom of the ninth, Riske gave up one run on a sacrifice fly by Chris Robinson. Later that inning, a run came in when James Adduci hit a sacrifice fly.