Envest

Portfolio Dashboard
Modal
Company List

Inspiration

Sustainability is no longer a nice-to-have, it’s essential for future-proofing investments and businesses. Investors struggle to find companies that align with their sustainability goals, while companies face challenges identifying best practices in environmental responsibility. Our team saw the need for a solution that helps both investors and businesses thrive in the sustainable economy.

What it does

Our platform helps investors and companies alike by:

Empowering Investors: Find investment opportunities in companies with strong sustainability practices using our advanced semantic search.
Predictive Risk Modeling: Provides future environmental risk scores for companies, helping investors build eco-friendly portfolios.
Portfolio Optimization: Investors receive personalized recommendations on how to improve their sustainability scores by diversifying into companies with growing eco-friendly initiatives.
Research Tool for Companies: Businesses can explore the sustainability strategies of industry leaders, using our platform to research and improve their own practices.

How we built it

We utilized cutting-edge technology to bring this vision to life. For the semantic search, we used a vector database (SingleStore) and cosine similarity to allow investors to search for companies based on specific sustainability traits. We integrated Hugging Face models to generate embeddings that power the semantic search. On the data analysis side, we used NumPy and scikit-learn to implement linear regression, predicting future environmental risk scores for companies. The environmental risk predictions and comparisons with the S&P 500 are visualized through Chart.js, giving users clear and insightful data representations.

Challenges we ran into

Building an app with this level of complexity came with its challenges. Implementing accurate semantic search required careful tuning of the embeddings and vector search algorithms. Collecting and standardizing environmental data across a large number of companies was also a big hurdle. Additionally, fitting linear regression models to predict future risk scores based on historical data required rigorous testing and validation to ensure accuracy.

Accomplishments that we're proud of

We are proud of the seamless integration of advanced technologies like semantic search and predictive analytics. Our vector search engine allows investors and sustainability teams to find relevant companies with incredible ease. Our linear regression models give forward-looking insights, helping investors and companies make data-driven decisions to improve sustainability. Most importantly, the potential for this tool to drive real-world environmental impact makes this project truly rewarding.

What we learned

Throughout this project, we learned the power of combining machine learning with sustainable business practices. We deepened our understanding of semantic search algorithms, vector databases, and regression modeling for predictive analytics. We also realized how crucial well-organized data is to create meaningful predictions and insights. Building a product that bridges sustainability with investment strategy was complex, but it gave us experience in balancing technical precision with user-friendly design.

What's next for Envest

Our next step is to expand the platform by incorporating more dynamic data sources, including real-time environmental news and company sustainability reports. We want to broaden our recommendation system, providing even more refined suggestions for portfolio improvement. On the company side, we plan to introduce tools that allow businesses to simulate the impact of implementing specific sustainability practices, helping them visualize potential gains in environmental scores.