M-Growth

Inspiration

With climate change and social inequality becoming increasingly pressing issues, we believe it's crucial for consumers to make responsible choices while shopping. Traditional methods of doing this usually involve a lot of research and time, making it hard for the average person to regularly make sustainable choices. We were inspired to build M-Growth to simplify this process, using data analytics and AI to offer personalized product recommendations that align with a user's values and needs.

What it does

M-Growth is a mobile and web application that aims to integrate directly into Migros's existing shopping platforms. When users shop online or planning their purchases by making a shopping list or looking up a recipe, they can use M-Growth to make more sustainably informed shopping decision, while staying in their budget.

How we built it

We used Python to process data and LLM models to Interprete the receipt and give recommendations.

To build a UI we used Streamlit library

Challenges we ran into

Integrating backend and frontend Waiting times API calls to models taking too long Vectorizing the German Data

Accomplishments that we're proud of

UI and data processing went smooth

What we learned

We learned a lot about the intricacies of recommendation systems and the importance of data preprocessing. We also gained insights into creating applications that not only serve a commercial purpose but also contribute positively to society and the environment.

Future of M-Growth

M-Growth recommendation system and the user journey can easily be integrated into existing Migros digital products. It offers both a pre-shopping solution for plaining sustainably for every budget and post-shopping solutions to reinforce this behaviour by rewards and labels for the customers.

During HackZurich23 we concentrated more on the pre-shopping part with customized recommendation systems for each customers. In the future one can improve the UX, offer more solutions and rewards for post-shopping and even consider implementing solutions for during-shopping.

Built With

• python
• streamlit