Wingmates

Inspiration

We, like many families in the U.S., have relatives living abroad. While our relatives have been safely able to navigate the airline systems so far, it has always been hard for them to adjust to all the different cultures they experience along their journey. Seeing our family go through this, we felt there needed to be a solution to make flying a more inclusive experience for international fliers.

What it does

To give international fliers a little bit more comfort as they navigate airports that host a wide range of cultural identities, we propose a social platform that allows passengers to connect with other fliers who share similar cultural backgrounds, as well as to view general flight information.

How we built it

Using React Native, Expo, and Render, we present a high quality application with a minimalistic--yet robust--UI to allow users to register and authenticate their login credentials through Firebase, view their current and upcoming flight information, and discover other passengers with similar cultural backgrounds.

The passenger matching algorithm drew upon principles of machine learning and natural language processing such as skip-gram word embeddings and continuous bag of words, we were able to pull cultural identity information from our user base in order to create a "culture vector" that we then compare with other passengers on the same plane using cosine similarity. The NLP was run using MATLAB and Python.

Challenges we ran into

Our limited bandwidth on Render constrained how many concurrent users we could test at once. Furthermore, the newest version of firebase had few examples of updated projects for us to learn how the API works, so we had a lot of trouble interpreting the stack traces with errors.

Accomplishments that we're proud of

We are very proud of our UI. We spent a lot of time planning it out on Figma and we were able to replicate it very closely. The real time data base, too, was very cool since we got to host users' information on our own Mongo and Firebase databases. Also, the matching algorithm has a lot of potential outside of airport interactions, and a lot of optimization led to it being a very fast algorithm.

What we learned

We learned a lot about the challenges in transforming an idea from paper to reality. There is a lot of planning, testing, iteration, and modifications that go into each product, and our app was no exception. We look forward to working and learning more about the technologies that we used in this project.

What's next for Untitled

We want to go ahead and find a way to connect multiple users through a chat that is maintained by a real time database in Firebase. Also, we want to possibly encourage the creating of groups of people that can translate for each other.