🌱 Sustainability Index – Fashion Industry Trends & Predictive Modeling 📌 Overview

This project analyzes sustainability trends in the global fashion industry from 2010–2024 using a dataset of 5,000 fashion brands. The goal was to uncover key drivers of sustainability ratings, identify opportunities for improvement, and predict ratings using a Random Forest model. Insights were visualized in Tableau dashboards to highlight environmental performance across materials, countries, and certifications.

🎯 Key Objectives

Analyze carbon emissions, water usage, waste production, material adoption, and certifications.

Predict Sustainability Ratings (A–D) using ML regression models.

Visualize trends and KPIs to enable data-driven sustainability strategies.

🛠 Tech Stack

Languages & Libraries: Python, Pandas, NumPy, Scikit-learn, Seaborn, Matplotlib

Visualization: Tableau

Data Source: Kaggle – Sustainable Fashion Trends 2024

📊 Methodology

Data Cleaning & Preprocessing

Handled missing & erroneous values

Encoded categorical features (Material Type, Country)

Created derived features (e.g., Average Carbon by Country)

Feature Engineering & Analysis

Correlation analysis between environmental metrics & ratings

Trend analysis for water, waste, and carbon by year and country

Modeling

Built Random Forest Regressor to predict ratings

RMSE achieved: ~1.17

Interpreted feature importance to identify key rating drivers

Visualization

Tableau dashboards showing yearly sustainability trends, material adoption, and certification impact

🚀 Results & Insights

Top materials: Recycled Polyester, Hemp, and Vegan Leather

High-emission countries: China & USA – need aggressive carbon reduction strategies

Certifications matter: Brands with GOTS or Fair Trade scored higher sustainability ratings

Weak correlation between environmental metrics & ratings → Suggests broader factors like labor practices also influence scores