As an solo international traveler, it can be daunting thinking about who we are going to be sat next to on the airplane. For instance, during the dinner meal of my flight out here, as I looked around, I saw many people chatting and enjoying sharing this dining experience with someone else. I had been sat next to no one, but would have vastly preferred to chat with someone. Our idea was pitched a few weeks ago, under the suspicion that it was ABD (already been done) by an existing airline. What we've done this weekend demonstrates how our approach is much better.

What it does

We integrate data collected from LinkedIn Profiles to readjust how passengers are assigned seats on the airplane. Using IBM Watson Bluemix, we compute some personal traits from the writing samples from passengers LinkedIn profile. We then compute a similarity score between each passenger using these traits and other explicit ones from their profile. Once we match passengers, we give them an opportunity to meet their match via private chat. If they like each other, they can confirm the match, check into the flight, and receive their seating assignment. If they don't like their match, they can disconnect and receive a random seating partner.

How we built it

We first thoroughly designed the overall system, analyzing many of the use cases of the feature. We then split up the roles of implementation: one person to create the database and our API, another to work directly with the data and assigning seats, and another to design and implement the UI/UX. We spent almost all the time at the venue working on the project, we each of us only getting a few hours of sleep this past night.

Challenges I ran into

We began to run out of time near the end as we were adding some of the features. We spent a bunch of time designing the backend database and API and run into problems with the actual deployment, specifically not having enough control with the first platform we deployed to. Both deployment schemes (heroku and aws) didn't smoothly deploy our application. Additionally, we face an array of challenges with user privacy and creating a smooth pain-less experience for users.

Accomplishments that we are proud of

We have an application that incorporates many ideas. We have a backend API, chatting feature, multiview webpage, and exciting data analysis.

What we learned

We might spend a bit less time designing the implementation of our internal API next time. While it is definitely crucial to have a fully functional backend, the time to implement one can be variable and in our case this time, it was a large portion of our hackathon time.

What's next for Finnder

We still have a bunch of options we would love to see incorporated into the application. We hope that it can be developed along side Finnair's ticketing process and eventually ends up in general deployment.

Built With

Share this project: