Inspiration

Carbon IQ is an AI-powered carbon impact dashboard. Users input lifestyle details — such as travel, food consumption, and energy use — and receive real-time carbon footprint estimates.

What sets Carbon IQ apart is that we:

  • Visualize data interactively using modern UI components
  • Offer personalized recommendations generated using AI
  • Compare a user's emissions to peers and global averages

How we built it

We started by preprocessing a custom emissions dataset using Python, pandas, and NumPy. After exploring feature relationships with a correlation heatmap, we used CarbonEmission as our target variable and split the data into training and testing sets.

We first implemented a baseline Lasso Regression model, tuning the alpha parameter with GridSearchCV. To improve accuracy, we moved to a Gradient Boosting Regressor, optimizing key hyperparameters (like learning rate and tree depth) using a Repeated K-Fold strategy.

Once our model was trained and saved, we built a RESTful API using Flask to serve predictions and insights on one's carbon's footprints and responses. On the frontend, we used Next.js and TailwindCSS, hosted on Vercel, while the backend was hosted on Heroku.

Challenges we ran into

  • Balancing the complexity of real-world data with simplicity of user input
  • Making the model performant enough for real-time predictions
  • Version dependency issues when hosting
  • Integrating the frontend and backend smoothly
  • Differentiating our app from existing calculators in a meaningful way

Accomplishments that we're proud of

  • A polished, modern UI built from scratch
  • Seamless full-stack deployment with a working API
  • Building something educational, beautiful, and extensible

Going beyond what’s already out there by adding comparisons and dynamic suggestions with pandas.

What we learned

  • How to combine ML + web development in a production-like environment
  • How to fine-tune models for usability and accessibility
  • The value of design in making technical tools user-friendly

What's next for Carbon IQ

  • Add login and user profiles to track emissions over time
  • Create a browser extension for real-time carbon feedback
  • Partner with schools or companies for group dashboards
  • Build integrations with smart devices for more accurate tracking
  • Open-source the API for developers and data scientists

Built With

Share this project:

Updates