AMADEUS, API, PROJECT
The backend is written in Python 3 using Flask and runs on a fun mix of Amazon Lambda, S3 and the excellent Python Anywhere. It’s something that I’m keen on developing - one of the most interesting things I’ve found is the amount of data that can be retrieved from a simple flight query that you would need to take into account: Things such as operating vs. marketing Airlines, Routing, Stopovers, Pricing Rules etc etc. I found that to make this app though I was only scratching the surface of what could be done with all this information and throwing a lot of valuable data away.
The pricing model of the data supplier (Amadeus) is that you pay per search performed, and so if you were to already to present to the user a predection that prices might be unfavourable - based on data that we’ve already seen, or that the ‘short hop’ actually consisted of 3 segements with a 4 hour connection time, you could both eliminate these results and hence expense, and also improve the user experience.
Much of the work was in the backend coming up with a simple algorithm for sorting through the results and trying to pair up flight itineraries that were sorted not only on cost, but the amount of ‘dead minutes’ the travellers would have to spend - i.e waiting for each other at the destination both arriving and returning.
It is currently running with a limited test dataset.