Inspiration
CO₂DE Red was created from a shared vision among a team with incredibly diverse backgrounds. Despite our different areas of expertise, we all agreed on one thing—we wanted to create something meaningful and impactful. We wanted our work to have an impact beyond the Biztech Tri-Mentorship program, something that could help people make more informed, sustainable choices.
A big part of our inspiration also came from our mentor, Riddhi Battu. She had extensive domain knowledge about the intersection of data science and sustainability, and we wanted to capitalize on this wealth of knowledge. We saw an opportunity to create something aligned with her expertise while pushing ourselves to build something relevant, especially as the widespread use of Large Language Models (LLMs) has grown. These models, while revolutionary, come with their own environmental cost. Training and running LLMs require massive computational resources, which in turn result in significant CO2 emissions. By addressing the issue of transportation emissions in our project, we were also mindful of the environmental impact of the technologies we were leveraging.
Because this project was created for an accelerated 3-week program, we needed to focus on a specific aspect of carbon emission production to make our project both challenging and achievable. We decided to concentrate on transportation emissions, the top contributor to BC residents' carbon footprints. This focus enabled us to create a tool that could meaningfully impact people's decisions while remaining achievable within the limited timeframe.
Riddhi's insights helped shape the direction of CO₂DE Red, ensuring that it was not only a practical tool but also one that could exist harmoniously within the broader conversation on sustainability in tech. We wanted our project to remind us that while innovation drives progress, it’s crucial to consider its environmental impact and find ways to reduce it.
Another key influence was the kickoff event for BizTech’s TriMentorship program. Someone said, “It’s one thing to identify the problem, but it's another thing to do nothing about it.” That stuck with us. We didn’t just want to highlight the problem of transportation emissions, we wanted to encourage people to take action. That’s why we incorporated Google Maps' API to make it easy for users to see their impact and have the opportunity to make more environmentally-conscious choices. By providing users with quantifiable measures and actionable insights, we hope CO₂DE Red red inspires real change, one trip at a time.
What it does
CO₂DE Red is a transportation emissions dashboard tracker designed to help BC residents understand and reduce their carbon footprint. Users input details about their trip, such as start and end points (by entering a street address or selecting points on an interactive map), fuel type, and transportation mode (e.g., sedan, minivan, transit), and the dashboard calculates the carbon emissions for the journey. For public transit, CO₂DE Red breaks down emissions, distance, and travel time for each step or mode of the trip. Additionally, the tool offers personalized recommendations to help users reduce their emissions and make more eco-conscious travel choices.
How we built it
Our tech stack primarily used the Streamlit library in Python, Google Maps API, and Google Gemini API. We used the Maps API to estimate the distance between the user's selected origin and destination and determine their travel route based on their chosen mode of transport. Using this distance and data from the BC government website, we calculated the average emissions for each trip. The Gemini API generated AI-powered recommendations to help users reduce their carbon emissions. Finally, we used Streamlit to display all this information in a dashboard.
Challenges we ran into
API Integration: Though extracting data from the APIs was straightforward, integrating it with the dashboard proved challenging and time-consuming.
Streamlit's UX Limitations: Working with Streamlit's restricted layout and visual options required creative solutions to build an effective dashboard.
Simplifying Complex Emission Calculations: We are dealing with complex data libraries and translating them into user-friendly metrics without oversimplifying or introducing inaccuracies.
AI-Generated Recommendations: We ensured that the Gemini API generated useful, personalized, and accurate sustainability recommendations. Avoiding generic or irrelevant suggestions that do not fit the user’s specific journey.
What we learned
Data-Driven Environmental Impact: By leveraging the BC Government's traffic database for emission factors, we created a solution grounded in real-world data rather than estimates. This demonstrates how government open data can power practical applications for public benefit. Integration of Multiple Technologies: The project showcases how diverse technologies (Streamlit, Folium maps, Google's geocoding, and AI feedback) can be seamlessly combined to create a cohesive user experience that addresses a complex problem. Translating Complex Science into Actionable Insights: We successfully translated complex data libraries into simple, understandable metrics that help users make informed decisions about their environmental impact. Personalization Through AI: Incorporating OpenAI's API demonstrates how AI can transform raw emissions data into personalized, actionable recommendations that feel relevant to each user's specific journey. User Experience First: The implementation of interactive map interface shows that simplicity in design streamlines the user experience making it available to everyone. Local Solutions to Global Problems: By focusing on BC-specific vehicle emissions data, we created a regionally relevant tool that addresses climate change at a local level while contributing to global carbon reduction efforts. Rapid Prototyping Potential: Using Streamlit allowed for quick development of a functioning application.
What's next for CO₂DE Red
In terms of future iterations, we hope to expand CO₂DE Red by incorporating other significant sources of CO2 emissions beyond just transportation. We hope to add categories such as energy consumption in homes, waste production, and even lifestyle choices, giving users a more comprehensive view of their carbon footprint. This would help individuals make informed decisions regarding their overall environmental impact.
We also plan to refine the user experience, perhaps by integrating additional data sources for more accurate emissions calculations, as well as enhancing the interactivity of the map and providing even more personalized, actionable recommendations. Another goal is to incorporate persistence and gamification or incentives to encourage users to track and reduce their emissions over time, making sustainability a more engaging and rewarding experience.
Built With
- gemini
- google-maps
- python
- streamlit
Log in or sign up for Devpost to join the conversation.