TEAM SODEVAN- FlyNet
We’re all wasting time at airports, bored in airplanes. It becomes difficult to establish a conversation with the one you’re travelling with. We want to create a connect, a relation between co-passengers which would benefit both of them in terms of business. We want to create a better experience for people travelling Vistara.
Benefits of Networking
Networking in aircrafts can be really effective because :
1)Confined space 2)You have to be there for a fixed amount of time 3)You have only one person near you to interact. 4)The diverse nature of people who visit airport serves as an added advantage for like minded people of different regions to interact with each other. However there is no such platform for actual interaction of such people.
Value Capture ::
1)Our product can be sold to various airlines. 2)Business Class customers would have a happy experience and would love to travel again by Vistara.
Exclusive Seat Selection using neural networks for Business Class Users
WHY BUSINESS CLASS?
1)Business class people are usually authentic people who are travelling for business work 2)They are regular fliers. 3)Would appreciate interaction for growing network with like minded people. 4)Grow network for Self interest/Business Needs/ Mentorship.
How have we done it?
1)Collect Professional User data from LinkedIn . 2)When User checks In, He is asked to select the Seat or Let us choose it for them. 3)If User let’s us choose the seat, We would ask him What’s on his mind. 4)This gives us the recent things up with him because Linkedin Provides A broad range data about a person. 5)Now we take keywords from the User Data collected from Linkedin and make a pool of all words of all users. 6)We make features out of those words and Make vector spaces for all Travellers. 7)Using Cosine law, we calculate the similarity between the profiles of Travellers. 8)The features of profile are passed into a Artificial Neural Network with changing weights so that importance of each feature changes accordingly. 9)The Neural Network clubs two people together based on the similarity score of the profiles and the What’s up status. 10)After the flight, a feedback system serves as the cost function for Backpropogation of Neural Network. 11)The weights are updated and the neural network improves after every usage.