Team No: 24//Room No: 2.05
Inspiration: We wanted to make the communication between customers and airline companies better. As a frequent flyer, we believed that for those that are more adventurous: they would have like some sort of roulette game where the app would choose their future destination. To do this we started by wanting the user to have some fun with the airlines application through a few games that would then lead to answers the app could use to analyse the users needs and thus find an appropriate city as a destination. We also wanted to make this more interesting by introducing a voice recognition system that would help plan the trip.
What it does: Our idea was to build a website that would allow a user to either have a conversation with a friend and let the chatbot plan your trip through the use of voice recognition or complete a form with 3 different areas of interest (weather, activity, population density) which would lead to an outputted list of destination suggestions.
How we built it: We started off by extracting 100 city destinations and their population densities from a Wikipedia. This would be our main data. With this data we used Python to generate a program that would classify each city depending on the three categories and output a “optimised” list of destinations that fit the users inputs.
We moved on to create a chat room using sockets as the middleware between the user and the voice recognition software. Using machine learning in DialogFlow, we were able to train the software so to recognise keywords such as cities, countries, dates, amount in groups.