How I Built a Real-Time ESG Risk Scorer using Python, NLP & Streamlit

As a final-year Computer Science student, I wanted to build something meaningful that combines AI and finance. That's how I built the ESG Risk Scorer — a real-time platform that analyzes Environmental, Social, and Governance risks for companies using NLP and live financial news.

What is ESG?

ESG stands for Environmental, Social, and Governance — three key factors investors use to evaluate a company's ethical impact and sustainability.

Tech Stack

• Python

• Streamlit (for the interactive dashboard)

• FinBERT (NLP model for financial sentiment analysis)

• Sentiment Analysis

• Real-time news data

How It Works

• Collects live financial news about companies

• Runs FinBERT-based sentiment analysis on each news item

• Categorizes risks into E, S, and G buckets

• Generates a final ESG risk score per company

• Displays everything on an interactive Streamlit dashboard Challenges Faced

  The hardest part was fine-tuning the NLP model to correctly classify financial news into ESG categories. Financial language is very specific, and generic sentiment models don't work well. Using FinBERT — a BERT model trained on financial text — solved this significantly.

  Live Demo & GitHub

  • 🚀 Live App: https://esg-risk-scorer.streamlit.app/

  • 💻 GitHub: https://github.com/2200032705-SREEJA/esg-risk-scorer

  Key Learnings

  • How to use domain-specific NLP models like FinBERT

  • Building and deploying real-time data pipelines

  • Creating interactive dashboards with Streamlit

  • Applying AI to solve real-world finance problems