FlightFlow

FlightFlow: Inteligencia visual que transforma cada charola y cada empleado en eficiencia máxima, llevando la operación de catering a otro nivel.

Comment

FlightFlow: Smart Intelligence for Airline Catering

Inspiration

Every flight is a microcosm of logistics, timing, and precision. With over 700,000,000+ passengers served yearly, airline catering demands flawless coordination.
We were inspired by one idea:

“What if computer vision could make airline catering as efficient as flight control itself?”

What it does

FlightFlow is a smart platform that merges data analytics and computer vision to boost airline catering efficiency across three key dimensions:

  1. Flights – visualize upcoming flights, operational details, and KPIs for accurate planning.
  2. Products – predict food consumption using a regression model based on flight type and passenger count, while detecting items close to expiration.
  3. Employees – apply real-time computer vision to measure tray assembly productivity and assign the best-performing staff to high-priority flights.

$$ \text{Data + Vision + KPIs} \Rightarrow \text{Operational Intelligence} $$

How we built it

  • Python (Pandas, Scikit-learn): for regression-based consumption prediction.
  • YOLO (You Only Look Once): used to detect, track, and analyze employee movements in real time, allowing accurate measurement of productivity and task completion speed.
  • React + Tailwind: for an intuitive, KPI-driven dashboard.
  • FastAPI: for backend integration and data flow.
  • Synthetic Flight Data: with realistic IATA codes, passenger counts, and stock cycles.

Challenges we ran into

One of the main challenges we faced was during the creation of our models, since the amount and format of the data we had were not exactly what we needed or expected.

Accomplishments that we're proud of

We’re proud of the ideas we came up with and how we executed and integrated them. We had excellent teamwork and knew how to organize our tasks efficiently.

What we learned

Beyond all the computing and programming skills we gained during development, we learned how important it is to identify the small gaps that often go unnoticed, those areas where processes are still manual and there’s a huge opportunity for innovation.

What's next for FlightFlow

This project is highly scalable with the potential to be trained on much larger datasets.
The video-counting system can be applied to various stages of the warehouse process, not just employee training.
Additionally, we plan to implement an AI-powered chatbot to create more direct and human interactions with administrators.

Note about our demo video

Unfortunately, our YouTube link was not accepted by Devpost.
We invite you to check the “Links” section on our project page to watch our demo video and see FlightFlow in action.

Share this project:

Updates