Spaces:
Sleeping
Sleeping
File size: 2,358 Bytes
2fbcfac 5600354 2fbcfac 5600354 2fbcfac 5600354 2fbcfac c2ae316 9b4bd5d c2ae316 9b4bd5d c2ae316 9b4bd5d c2ae316 9b4bd5d c2ae316 9b4bd5d c2ae316 9b4bd5d c2ae316 9b4bd5d c2ae316 9b4bd5d 2fbcfac |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 |
---
title: GreenPrint AI
emoji: 🚀
colorFrom: red
colorTo: red
sdk: docker
app_port: 8501
tags:
- streamlit
pinned: false
short_description: A futuristic take on carbon footprint + AI.
---
# 🌿 GreenPrint AI
> A futuristic take on carbon footprint + AI.
## Overview
**GreenPrint AI** is a user-friendly web app that predicts your **carbon footprint** based on daily activities and suggests actionable steps to reduce it. From energy consumption to travel and food habits, the app personalizes insights using machine learning.
---
## Aim
To develop a carbon footprint detection system that:
- Takes input as user activities (e.g., electricity use, transport habits, meat consumption).
- Predicts the carbon footprint in kilograms of CO₂ equivalent.
- Recommends practical, low-carbon alternatives.
---
## Step-by-Step Project Workflow
### Step 1: Dataset Creation
- A synthetic dataset named `synthetic_carbon_footprint.csv` was generated using realistic formulas for CO₂ emissions from:
- **Electricity consumption**
- **Travel (personal & public)**
- **Dietary choices (meat)**
- **Plastic usage**
- These were computed using coefficients (e.g., kg CO₂ per kWh, km, gram of meat).
- The script `Creating_Dataset.ipynb` contains the exact formula-based generation logic.
---
### Step 2: Model Training
- Model Used: RandomForestRegressor from scikit-learn
- Training done using `Running_Model.ipynb`
- Preprocessing involved:
- Feature scaling (`StandardScaler`)
- Ensuring input columns align with model_columns.pkl
- Splitting dataset into training/testing sets
- Evaluation metrics:
- **R² Score** for goodness-of-fit
- **Root Mean Squared Error (RMSE)** for prediction accuracy
- The trained model was saved as `carbon_model.pkl`
### Why Random Forest Regressor?
- It handles non-linear relationships and interactions between features better than linear models.
- It’s robust to overfitting due to its ensemble nature.
- Outperformed linear models in testing, giving: Higher R² and Lower RMSE
- Offers feature importance insights, useful for explainability in a sustainability context.
---
### Step 3: Feature Metadata Handling
- A `model_columns.pkl` file was created to store the expected feature column order.
- This prevents column mismatch during inference in the Streamlit frontend.
---
|