The Tnooz THack event at ITA Google's headquarters in Boston was a lesson in how overwhelming development for the travel industry can be. On the morning of the first day of the hackathon, the eleven API hosts provided brief descriptions of their APIs and their respective functionality or access to data. It became clear, very quickly, that most of the APIs were not like the Twitter API or Google Maps API. Many of the APIs delivered a lot of complex data in a highly structured format. Building some kind of mash-up using any of the APIs would be a challenge.
The first hack, presented by the Amadeus team combined the Amadeus hotel API along with flights, the Silver Rail API for rail tickets, the Goby API for things to do, along with the Facebook API for social graph. The purpose of the hack was to allow a combination of social trip planning with dynamic package building around a user's favorite NFL team. The dates and locations of the games were hard coded into the app (or in a final version in the database). The flights, hotels, and activities were then triggered based on the dates and locations set with the chosen football game. Despite the limited time frame, the application managed to allow for searching and retrieving content from the various APIs, and building a basic itinerary along with the ability to share the itinerary with friends.
The second hack presented by the developer from Delta Airlines was a partially working Android application that used stubs but didn't connect live to the APIs. The application ran as a widget on the user's phone and pushed content to the user in real-time based on their location and any current upcoming trips. The hack used the Goby API for location based activities and the Traxo API for trip planning purposes. When asked about possible features, the developer talked about being able to push deals to the mobile user in destination as well as being able to notify the user of real-time updates from their itinerary based on delays or changes in their trip.
The third hack was by a developer with a background in hotel market analytics. The hack focused on providing the traveler with the ability to track and compare average hotel pricing for a specific trip compared with the activities that are available near the destination. The Amadeus and Goby APIs were used to pull hotel live pricing based on location. He would have liked to have added airfare and weather to the mix in order to provide more decision factors to the process. The developer also mentioned that because of the way the Amadeus API was written, it required him to do 180 separate XML requests in order to retrieve 6 months worth of hotel pricing data, for trending purposes.
Given the number of APIs, I would have expected more of the APIs to be used. So, why did the Amadeus and Goby APIs shine over the others? I think the short time frame and limited programming resources had a very important role in the API decision making process. It also demonstrated one of the most frequent complaints from developers who want to work with travel APIs in general and that is that the APIs, the data, and the processes required to put even a simple application together are too complex. Searching for and booking a hotel, for example, is a much more sophisticated process than retrieving tweets from a user account.
The beauty of events such as Thack is that it sheds some light on the capabilities and limitations of currently available travel APIs and also generates interest in both the development community and the application providers themselves. Couple that interest with the need to create some standard messaging and we could see a growing number of APIs coming on stream as well as an increase in development around travel APIs.