Inspiration
We noticed that in events, hotels, colleges and functions, food is often prepared based on the expected number of people, but the actual attendance and food consumption can be very different. This can lead to a lot of unnecessary food waste.
So we thought: What if we could predict how much food is actually needed, and if there is still safe surplus food, help get it to people who need it?
This inspired us to build Kibo Meals. And also I wanted to thank for conducting dev post zero origin hackathon and giving us this opportunity to implement our kibo meals website.
What it does
Kibo Meals is an AI-powered food waste management platform that considers factors like attendance, event type, weather, holidays, meal type and previous consumption to predict the required quantity of food.
After an event, the system can track how much food was prepared, consumed and left over. If safe surplus food is available, Kibo Meals can help identify nearby NGOs, old-age homes, child-care homes and community organizations that may be able to use it.
Our workflow is:
Predict -> Prepare -> Track-> Match -> Redistribute -> Learn
How we built it
We created a synthetic dataset containing event, attendance, food, weather and waste information. We use Python, Pandas, NumPy and Scikit-learn for the data and machine learning part, with HTML, CSS, JavaScript and Flask/FastAPI for the web application.
The website will provide demand predictions, food waste analytics and a way to find suitable recipients for safe surplus food.
Challenges we ran into
Our biggest challenge was not having a large real-world dataset containing all the information we needed, so we started with a realistic synthetic dataset.
Another challenge was handling unpredictable attendance and making sure that only safe surplus food is considered for redistribution.
Accomplishments that we're proud of
We are proud that we were able to turn a simple idea about food waste into a complete concept that combines AI-based demand prediction, waste tracking and surplus food redistribution.
We also created a realistic synthetic dataset and designed the workflow from predicting food requirements to finding potential recipients for safe surplus food.
Most importantly, we are proud that our project focuses not only on reducing waste, but also on turning extra food into something useful for people in need.
What we learned
We learned how different factors like attendance, weather, holidays, events and food preferences can affect food demand.
We also learned about preparing datasets, machine learning, building a web application and designing a system around a real-world problem.
Most importantly, we learned that solving food waste requires more than just prediction — it also needs tracking, decision-making and proper redistribution.
What's next for Enjinia
Our next goal is to improve the prediction model using real-world data and make the recommendations more accurate. We also want to add real-time weather and location services, a better recipient-matching system and notifications for surplus food. In the future, Kibo Meals could be connected with restaurants, caterers, hotels, colleges, NGOs and community organizations to create a larger network where surplus food can reach the right people at the right time. Our long-term goal is to make Kibo Meals a platform where less food is wasted and more meals become hope.
Log in or sign up for Devpost to join the conversation.