Hacker News Reader

My problem was Twitter's API not having a more intelligent filter API. If you want to use it to filter all tweets with stocks (i.e. "\$\w{1,4}") you can't as far as I could tell. Instead you have to filter on '$' which gets you a ton of noise ("got me some $$$ for 2NITE! yeah!!!!").

View on HN

Twitter has an incredible primitive search. A search for C++ searches for C instead of recognizing the whole term.

... Which you could them pull into a database and filter with your Regex after the API. What does the noise cost you? It's a solvable problem. true, the expanded results may keep you from getting relevant tweets due to rate limiting.

But that is the point, you get rate limited. Of course you can take the data set and clean it in your own database. The problem is how much relevant data you can get before getting cut off.

Reply on news.ycombinator.com