Foodability

Start: Here you can enter the ingredients/dishes you like
Selection: Select all the recipes you like or remove any that you don't
Foodability gives you a shopping list and links to the recipes!

Inspiration

Food production is a significant contributor to climate change. Thus, it is particularly shameful that so much food is wasted (see e.g. https://www.bundesregierung.de/breg-en/news/halving-food-waste-1582700). Therefore, we want to support you to actively minimize food waste and reduce packaging.

What it does

You enter ingredients or dishes you like, and Foodability plans a cooking schedule for you. When buying ingredients for a specific recipe, it is often challenging to completely make use of all items. Hence, Foodability suggests recipes that complement each other, such that food waste and packaging is minimized. Additionally, we show you how much food is potentially wasted and illustrate the amount of food packaging.

How we built it

We analysed the Open Food Facts data (https://world.openfoodfacts.org/data) and engineered features that indicate the amount of packaging. Afterwards, we collected many recipes and normalized/harmonized the data. Once our data was in place, we designed an expectation-maximization-style algorithm that optimizes recipe suggestions with respect to food waste and the amount of packaging. Finally, everything comes together in our intuitive and responsive UI.

Challenges we ran into

As in many data-driven projects, data pre-processing and feature engineering were challenging aspects. Particularly, the pre-processing was time-intensive and different formats and units across datasets required careful handling. Further, we encountered some difficulties with the deployment of our application to Azure. We circumvented the problem by switching from Azure Containers to Azure App Service.

Accomplishments that we're proud of

We are proud that we managed to implement a working version of our ambitious idea. Considering all the complicated and involved tasks we had to solve, from data engineering to UI/Backend design, it is quite remarkable that it worked out so well. Of course, our application can still be improved, but we believe we have completed a solid piece of work. Also, despite the current pandemic situation, the communication in our team worked flawlessly. Something that does no go without saying!

What we learned

We learned that data cleaning and pre-processing require a lot of time and should not be underestimated. While we encountered difficulties, the clear communication in our team allowed us to resolve issues quickly. Therefore, the importance of communication is also a big learning for us. Lastly, because of the limited time frame, we also had to learn to make compromises.

What's next for Foodability

We hope that Foodability can give you a more conscious perspective on food waste and packaging. Given more time, we would like to further refine our data to give you more precise information. Also, metrics we have not considered yet could be very insightful. Last, but not least, we would like improve our application from a software engineering perspective. This includes making the code more readable, maintainable and easier to extend.