Timely Travel

Inspiration

When researching the subject, it seemed clear that one of the biggest pain points for airline customers is flight delays and cancellations. After collecting some data we were astonished at just how frequently and predictably flight schedules were disrupted. This presented an opportunity for us to take a crack at solving an age-old problem with cutting edge technology.

What it does

TimelyTravel is an app that helps travelers book flights with confidence. The app uses a predictive model trained on nine years of real-time data to provide accurate flight arrival and departure times. This helps travelers plan their trips and gives them peace of mind that they will get to their destination on time.

How we built it

We used flight data from American Airlines and Duffel which is consumed by our mobile client and reliability prediction service. We used a combination of Go, Javascript, and Python

Challenges we ran into

Managing incredibly large payloads from very picky API's was tough.

Accomplishments that we're proud of

Our prediction model exceeded our expectations.

What we learned

When entering this hackathon we were all faced with a new challenge: Andrew - Normally working solely on frontend was assigned to work on the backend helping expand his knowledge on Go. Raaj - Also switching it up, Raaj was placed on the frontend team expanding his knowledge with Vue3 while also assisting with other members. Jacob - Jacob was also assigned to the frontend team and was tasked with building a plethora of components. Given that this was Jacobs first time using Vue he was thoroughly challenged. Roman - Expanding his knowledge of ML, by deploying his first feed-forward neural network on flask.

As a whole we struggled quite a bit with connecting our multiple micro-services together however collective debugging we were able to solve most of our problems.

What's next for Timely Travel

A more recent dataset would have vastly improved our accuracy on daily flights given that our dataset only went up to 2018.

As for development, Timely Travel plans continuing to test and debug our application improving user experience and decreasing query times on our larger search routes.

link