What does it take to get the airline price and schedule data?
I imagine a lot of people want to make sites like this, but they don't have the resources to obtain contracts for data feeds.
Much like many people would like to open up streaming sites for Big Label music, but not everybody can afford to enter into contracts with music labels then have hundreds of HDs full of wav files delivered to them.
We need someone to make a pay-per-taste brokered data market.
Basically, getting the raw data is quite difficult, but once you have it, you need to expend a really enormous quantity of engineering work into using that data to get useful answers.
For example, ITA invented a language to describe taxes. That's because the tax code changes all the time and is incredibly complex. There are fares which different entities in the airline industry price differently due to tax code ambiguities. The search problem is really very hard once you have the data.
Now, ITA software does offer their flight search engine as a service: you sign a contract with them and issue queries and they send you responses. I don't know how interested they are in cutting deals with really tiny startups, but they might be amenable. You will of course have to pay in some fashion: maybe a fixed fee per query or pay per compute time or pay per successful booking.
That's very illuminating. I have some ideas for displaying / managing flight and travel information that I would really love to implement, but the "Data Problem" has loomed large for experimentation. I didn't in fact realize that it was even as burdensome as you describe.
Just issue query to Orbitz via the affiliate API and they'll pay you for every flight booked. The TOS have some limitation, but if you can work within them you can explore pretty easily.
It's really quite ridiculous. I made the backend to a travel site using data from Galileo reservation system (one of the big ones). At the time, they had a relatively new SOAP-ish API. The data you got back was huge and arranged in ways that are not in any way obvious. Many calls just return what appear to be screen scrapes of 70's era terminal screens, wrapped in XML. It was less than fantastic.
At the time I left the travel agency in question, Galileo was in the process of rolling out a new version of the API, but the few times I tried to get it to work, it seemed completely broken (this was for cruises, not flights, but it was the same general idea).
I imagine a lot of people want to make sites like this, but they don't have the resources to obtain contracts for data feeds.
Much like many people would like to open up streaming sites for Big Label music, but not everybody can afford to enter into contracts with music labels then have hundreds of HDs full of wav files delivered to them.
We need someone to make a pay-per-taste brokered data market.