EuroOSCON: Adrian Holovaty – Journalism via computer programming

Journalism right now is broken. Several ways – celebrity focus, political bias, circulation declining consistently, stock prices dropping, craisglist taking away classifieds market.

But that’s not the issue. The issues is that newspapers throw away data.

So if there was a burglary, you have the address, the person, the stuff nicked, roughly the time. has key value pairs. But all the journalist does is write an article and throw away most of the data.

News orgs have huge infrastructure, with reporters on the street, specialised. Infrastructure to collect and edit information, verify it. Not every media organisation does that, that they’re not taking advantage of. Have infrastructure to get info out to people, i.e. a printing press originally. Also have the attention of people.

But can’t take advantage of data because they are just creating big blobs – stories.

So contrast to
Google Base, (which is just infrastructure with no data).

All great frameworks desperate for data. Journalists have great data desperate for a framework.

Why is structured data important – because if it’s structured a computer can do cool stuff with it.

Journalism via computer programming.

News people write an article, or create a video. A programmer makes a web app that makes it easy to look at the data.

WaPo, Iraq war, huge issue. Most recent deaths page, total fatalities, in Operation Iraqi Freedom, or Operation Enduring Freedom. Collect data on everyone who’s died, but can’t do a story on everyone, but can make that data available.

Faces of the Fallen – get own page, bio, map of home town. Depressing but important. Breakdown of age of deaths, most are 21, look by age, photos, breakdown of state, see all the people from the state and their town. Googlemaps. RSS Feed for every state. Sounds depressing and gory, but people are interested and they are then making their own sites, using for political activism.

Another example:

Type of crime, street, by block, brows by day, by hour, and latest crime RSS feed

Votes Database, representatives in congress, their votes, breakdowns of late night votes, votes missed, get RSS out there. So people can get more interested in government: did you know your representative voted this way today?

Telling a story via an application not words. Being smart about data, dealing with raw data. Badger journalists to get the raw data so we can do cool stuff. End game is not creating an article, but getting data in one place to do cool stuff.

Cultural similarities to this and open source code.

Open source:
– making code available.
– understanding through transparency: can download stuff and look at it.
– encourages derivative work, although depends on licence.

Journalism via code:
– make the data available.
– encourages understanding through transparency: better to look at the data than someone’s opinion.
– encourages derivative work, can take the data from the RSS feed and do stuff with it yourself.

Call to action
Done talks at journalism conferences, and people grumble that this ‘isn’t journalism’, but that’s kinda depressing that the industry thinks that way. It’s not full of passionate people who want to do cool things with technology but full more of people more interested in the ends than the means. So if people are interested, then go out there and do it.

Comments are closed.