-
-
Start screen
-
Profile screen showing performance and badges
-
Bar code scanner in action
-
Example results
-
Example results
-
Example results
-
Example results
-
Example results
-
Example results
Inspiration
We wanted to make it as easy as possible for consumers to choose (more) sustainable products at Migros. For the reward system we were inspired by step counters and other fitness trackers.
What it does
The rECOmmender makes suggestions for similar but more sustainable products with regard to CO2 emissions. This can be done by scanning the barcode of a product using the app. To encourage the user to buy these sustainable products, there is a preview of how gamification might work based on badges.
How we built it
We have developed two separate components:
- A recommendation engine to generate sustainable alternatives for each product
- An Android app to provide a UI for the recommendations and to monitor your own performance and badge achievements
Challenges we ran into
Making good recommendations for alternative products requires an extensive and clean data baseline. Insights into the data structure and expert knowledge are critical for the success with the task. The latest versions of the build tools, stopped to work smoothly together as soon as we added libraries for the barcode scanner. It took our frontend-dev several hours to get it working.
Accomplishments that we're proud of
We managed to create a working prototype within a single day and to connect the recommender and the mobile application including the barcode scanner. The tool can allows users to make choices that are good for the future of our planet. And it even looks good 😉
What we learned
The hierarchy of the products of a retailer is a complex topic and there are no easy solutions. Additionally, setting up environments and getting libraries to work takes a non-negligible amount of time, but if you keep trying, you'll finally make it. You can be busy at a Hackathon all the time without coding a single line.
What's next for rECOmmender
- An improved recommendation engine: Due to modular structure of the codebase, one can aim to replace the code in the prototype with calls to the recommender systems that probably are available internally at Migros.
- Figuring out how to make gamification even more interesting
- Adding the 'Tierwohl' labels to the UI and including them in the suggestion computation
- Moving data out to servers instead of computing every thing locally
- "Make the planet green again!"
Log in or sign up for Devpost to join the conversation.