tldr: All websites from HN&co-outgoig links indexed and prepared for a machine learning task (not only "normally" indexed, but tokenized, POS tagged, Sentiment Analyzed, NER, ..., know how many comments, points, was on which site of HN, termVectors, etc.). What would you do? And what to do to earn money with that?<p>long story:
I was having some fun with financial-News-data and the Forex market for my masters thesis. Well: no, I did not get rich :). But I've learned a lot. A lot about Sentiment detection, about NLP, about all that stuff. I found bitcoins are a fun playground, but I do not want to put my money directly into the whole thing. I want to provide a service that could help cluster/filter the news in a very specific way so that someone who is really into the financial markets in any way could benefit from it - or something completely different. I am crawling constantly since Oct 2014 from HN/Reddit and 36 other datasources links and comments and I could do a lot of fun things. A neat sideproject resulting out of the whole thing is http://search.nagrgtr.com or another is a similarity service, where similar articles are found to an url in the index: http://nagrgtr.com/morelikethis . Nothing special, but still a pretty nice to have. The question is now: What do you think would be beneficial to do in order to earn a living out of it? A google-news for hackers? Or something completely different? Any ideas anyone?
I am also wondering of some companies have made money off these things.<p>I know of some examples off the top of my head, but I don't know too much specifics on how the idea came from or what exactly the sell/how they make money. I think a lot of the money comes from investors until they really start selling a product, and selling a product of just information is somewhat hard I think.<p>For example:<p>Mattermark-> tracks startups and provides info for VC firms and regular joes<p>Dataminr, StockTwits-> Monitor twitter for viral events or breaking news