The second Tnooz THack Bangalore showed not only how engaged the developer community is with travel tech, but also just how clever some of the ideas coming out of Bangalore are.
The one-man team TripBuddy was the star of the show, winning the judges' prize for the best small team and also the people's choice prize, voted for by the other developers.
TripBuddy was invented by Dikshant Shahi, a developer with a background in semantic and voice-based search.
The hack was a live chat platform which could be plugged into a web site or app, or even operate as a standalone messaging service. At its heart lies AIML - "artificial intelligence mark-up language."
Put simply, TripBuddy used AIML to convert the search query and results into natural language which would work within a chat environment.
The example used in the presentation was a simple flight query which connected with the Amadeus API to bring back the results. The beauty of the hack is that, theoretically, the idea of using AIML to translate search queries into natural language could work with any travel API.
Or at least, Dikshant explained, any travel API which has metadata structured in a way which can be read by AIML. On the day of the hack he realised that the flight data within Amadeus was compatible with the live chatbot idea he had been thinking about for some time and set to work.
The rapid turnaround of this idea into a workable demo within the time-frame impressed the judges as much as the demo itself.
For queries more detailed than simple flight availability, the travel APIs would need to be structured differently.
Dikshant is now about to do what all developers do, and go travelling! His interest in taking part in the hackathon was driven by his interest in travel as much as his interest in technology.
He has launched his own blog, www.hobotrails.com, to record his experiences. He will continue to work on a freelance basis and has a few good leads as a result of the hackathon.
Elsewhere at the hackathon, developers showed their creativity when it came to interpreting the five pre-defined challenges they were asked to address: inspirational search, mobile, multi-modal, personalization and wearables.
Noticeable by absence was a lack of interest in wearables, with none of the developers taking on this challenge. This might be because India is still "emerging" in terms of adoption, but it is equally possible that wearables are seen as a gimmick in a market where the core internet real estate of desktop and mobile still have so much to offer.
Taken out of context, "inspiration through personalization" sounds like the title of the world's most pompous self-help book, but it works in a travel tech context. It has long been the Holy Grail of travel tech to inspire - read "persuade" - travellers to book inventory or supply which has been personalized - read "targetted".
Personalization is the driving force behind NDC, behind merchandising, behind big data. The zeitgeist of the moment is all about sifting through the myriad of options to come up with only the results which are relevant.
This inspirational/personalization combi was the most accepted challenge in Bangalore
Jan-Paul (JP) Ephithite, Senior Product Manager from Travelport, one of the event sponsors with Amadeus and HolidayIQ, suggested that inspirational search tools could bring up just one, perfect result.
This was intended as a call to arms for the developers, not a statement of one of Travelport's corporate goals.
At the heart of many hacks was Facebook's social graph API. Pulling in data from Facebook, mining that data for keywords and then connecting with a travel supplier - more Holiday IQ's attractions database than Amadeus' flight inventory - was tried by a few of the teams.
One hack connected a travellers' calendar with Facebook, an attractions database and a transportation API and suggested things to could be visited between meetings.
Another hack pulled on events data from Facebook Events, Eventbrite and Meet-Up APIs to suggest smaller scale events for business and leisure travellers - promoting gigs in pubs over gigs in stadiums - as part of the "travel like a local" trend.
Localization wasn't flagged up as one of the challenges but many teams used location-based APIs - Google Maps the most popular by far - to produce itinerary-based products.
A few of the hacks stood out by their individuality - one team used the Flickr and Instagram APIs to create a heat-map to show how busy attractions are at different times of the day by referring to the timestamp on the images.
Another ambitious hack which fell short of being able to be demo-ed was based around trying to sell vacation rental by interrupting a hotel transaction with a pop-up during the booking flow.
Overall, twenty-one hacks were presented over a slightly chaotic three-hour session on the Sunday afternoon. One of the judges, Shekhar Kirani, a partner at investment fund Accel Ventures, praised everyone who had attended, saying that most had managed to achieve in 24 hours what many startups would take months to get to.
As a veteran of between 30-40 hackathon judging panel appearances, and as a VC who deals with 25 start-ups a week, Kirani gave the room some free advice - be a specialist and be able to explain your value proposition in 140 characters or less; highlight how different your product is to what is available in the market; and spend more time preparing for the actual presentation of the hack to the judges.
Anand Kandadai, former-vice president at Indian OTA MakeMyTrip also offered some free advice specific to travel - "think about who is the customer and how you are going to acquire them."
If there is one thing that the second Tnooz THack Bangalore shows, it is that the API economy is gaining momentum. Merging public APIs with travel-specific ones gives developers the chance to experiment with ideas, some of which work, some of which don't, some of which may end up as the Next Big Thing. The important thing is that APIs are fuelling these ideas, and so long as the ideas keep coming, travel will remain the most dynamic e-commerce sector there is.