Off the top of my head: -- Factual (http://www.factual.com/ http://www.factual.com/), who I really, really like because they want to develop an "open data plat

Off the top of my head: -- Factual (http://www.factual.com/ http://www.factual.com/), who I really, really like because they want to develop an "open data platform" -- Wolfram Alpha (http://products.wolframalpha.com/developers/ http://products.wolframalpha.com/developers/) licenses their super-amazing API -- Twitter (through Gnip) licenses their data to users (http://gnip.com/twitter http://gnip.com/twitter) Now, in my opinion this kind of business model is awesome for a few reasons: 1. It's not glamorous. No one says "whooooo I want to develop a data API and license it!" Everyone is too busy developing a Facebook for cats or whatever. 2. It's really, really useful. If you work with data in any capacity, you know that an API for said data is much better than scraping data from HTML. Even the HTML-parsing modules in Python/Ruby/Perl, which are friendly relative to raw HTML + regexes, are brutal. 3. It's super-interesting. The world has so much unstructured data, and yet this market is (almost) untapped! There is data on almost anything you can think of -- health care records, pedestrian/vehicular/retail traffic, food consumption -- and you'd only need some thought, maybe a few data-collecting tools, and an API to make it useful.