News for hackers. Stuff that matters.
Re:Introduction to Zipline: A Trading Library for Python by smnscu on Monday July 25, 2016 @05:06PM (#12161615)
Random idea, but what if you could combine market data with news articles, and (obviously only for the past few years) data from blogs/reddit/hn/twitter/etc.? Sometimes there is fascinating insight to be learned in obscure places on the Internet (hn itself being a great example); I imagine a system that would collect and analyze that kind of information would be quite interesting, despite the huge amount of noise. Such a hypothetical system reminds me of this story: https://www.facebook.com/notes/robin-sloan/julie-rubicon/985...
Re:Introduction to Zipline: A Trading Library for Python by chollida1 on Monday July 25, 2016 @05:09PM (#12161630)
Glib answer....

You'd get really mad at Anne Hathaway

http://www.theatlantic.com/technology/archive/2011/03/does-a...

Longer answer, this has been done for 10+ years. Bloomberg will sell you a sentiment annotated news feed for 5 figures a month if you'd like to try.

Re:Introduction to Zipline: A Trading Library for Python by mikkom on Monday July 25, 2016 @05:16PM (#12161669)
This has been made many, many times already. Zipline that we are talking about here is created by Quantopian that actually provides aggregated twitter harvesting datafeed free for this year.

https://www.quantopian.com/data/psychsignal/stocktwits

Re:Introduction to Zipline: A Trading Library for Python by alfalfasprout on Monday July 25, 2016 @05:33PM (#12161768)
Been done tons of times. Some of them make quite a bit of money, others not. The real game-changer is if you're effective at separating out the real signal from the noise (and the noise floor is high in these sorts of feeds).
Re:Introduction to Zipline: A Trading Library for Python by nl on Monday July 25, 2016 @11:33PM (#12163158)
Re:Introduction to Zipline: A Trading Library for Python by hiou on Monday July 25, 2016 @06:48PM (#12162162)
You would probably have as much luck correlating features like whether your neighbors kid was at the bus stop or had to run to catch it each day and the amount of times your neighbor complained about someone not cleaning up their dog's poop that week. Bullshit random edge case features don't get any better or worse if they come from the internet.

In regular trading(algorithmic or not) you really need some non-trivial insight that for some reason no one else will have.