News for hackers. Stuff that matters.
Re:Chicago's top crime blogger is a 16-year-old with autism ... by delackner on Wednesday August 26, 2009 @11:13PM (#788556)
Application idea: given that voice recognition systems are not up to the task of parsing a police scanner's output, dice that police radio feed's audio into a single chunk per "transmission" and add that audio to a mechanical turk work queue. The work task is simply to enter into a form: is this a crime? yes/no. location: street, street. text details: (free text).

it wouldn't be perfect, but could this work?

Re:Chicago's top crime blogger is a 16-year-old with autism ... by jim-greer on Thursday August 27, 2009 @04:15AM (#788865)
This would be cool but I think it would be hard for the average Mechanical Turker to get it right. They'd have to learn all the numeric codes police use for everything, for one thing. Also the dispatcher and cops know the local geography very well and use a lot of shorthand. Here's the live police scanner feed for Seattle (mostly silent right now since it's 1am):

http://www.komonews.com/news/content/scanner/31214074.html?v...

Lots of other feeds are linked here:

http://www.police-scanner.info/live-police-scanners.htm

Having said that, I don't think you need to hire mechanical turkers - you could make a nice app for the community of people who are really into listening to this stuff, and they'd do the work for you - Timmy is not alone. Make a community site around that does a nice job aggregating the feeds, and that rewards accurately reporting them and adds maps, trends, etc. Not sure there's a business there - you might be able to get some grant money to get it off the ground though.

Re:Chicago's top crime blogger is a 16-year-old with autism ... by madebylaw on Thursday August 27, 2009 @01:28AM (#788710)
I was thinking something similar. You could have police scanners placed around different cities to get everything. I didn't know whether voice recognition could parse the output though, it's not really my field.