The fact that our generations are dealing with an inmense environmental problem, aggravated by the food wastage happening everyday.
But not only is this an environmental issue, food suppliers are lossing over 750 billion dollars every year on food wastage.
We decided to contribute avoiding this problem and created a food demand forecasting application that will help workers and companies adjusting their purchase orders and thus avoiding over production and useless contamination.
What it does
It forecasts the demand for every type of product given a food supplier for the 10 upcoming weeks. It also helps setting optimal prices to assure all the stock is sold, hence avoid food wastage.
How we built it
We used python and many libraries, such as scikit-learn, XGBoost or matplotlib, in order to create several models in the search of the most accurate prediction.
Challenges we ran into
Give that we have no front-end developer in our team, the UX designer had to work really hard to create a complete usable mockup on figma.
We have had issues with the model's accuracy as well as with its size. The weight and limitions while deploying the model have been an usual constraint.
Accomplishments that we're proud of
We have a working deployed model and a glance of the real app through a figma design.
We have tested several models in search of the optimal algorithm and parameters
What we learned
We have improved our knowledge of time-series data and how to work with them. We have improved our collaboration techniques. We created a regular schedule and tasks assignment that led us to a fulfilled project under stressing time conditions. Deploying techniques and usual errors that come with it. The importance of overcomunication.
What's next for Less is enough
- Converting the figma mockup into a usable front-end application.
- Creating a AI model based application that can improve our results.
- adding new features such as recommended order.