Here's an idea for next time. Assuming you don't need the data in real time, but instead just need to be able to say, when X story broke, we broke it first 80% of the time.
This is probably a stupid idea but I've only given it a few minutes of thought, so thats all you get tonight :)
Take automatic snapshots of news pages, like Drudge.com, cnn.com or whatever you want to compare to. Maybe once a minute.
Now hook it up with Mechanical turk and pose the question (in turker language its HIT (human intelligence task), "does X story appear on this page at 10:00pm?". They answer as true or false, then keep moving the capture forward and keep asking the question.
Comments
Here's an idea for next time. Assuming you don't need the data in real time, but instead just need to be able to say, when X story broke, we broke it first 80% of the time.
This is probably a stupid idea but I've only given it a few minutes of thought, so thats all you get tonight :)
Take automatic snapshots of news pages, like Drudge.com, cnn.com or whatever you want to compare to. Maybe once a minute.
Now hook it up with Mechanical turk and pose the question (in turker language its HIT (human intelligence task), "does X story appear on this page at 10:00pm?". They answer as true or false, then keep moving the capture forward and keep asking the question.