Atharwa27/Risk_analysis

⚠️ AI PDF Risk Analyzer

AI-Powered PDF Risk Analyzer is a modern, Streamlit-based web application designed to help professionals analyze legal or business documents for potential risks. The system leverages a custom word-risk dataset and AI embeddings to identify risky terms and calculate a document’s overall risk level.

Project Overview

• —Purpose: Quickly assess PDF documents for high-risk terms and quantify the document’s risk level.
• —Technology Stack: Python, Streamlit, Sentence Transformers, Plotly, Pandas, PDFPlumber.
• —Key Features:
• —Upload PDF files and extract text efficiently.
• —Detect risky words based on a custom CSV dataset containing words and risk scores.
• —Compute average risk and risk contribution per word.
• —Ignore stopwords and punctuation to improve accuracy.
• —Visualize results with modern, interactive bar and pie charts.
• —Dark theme with purple gradient cards for a professional look.
• —Real-time progress bar for processing large documents.

Highlights

• —Interactive Analysis: Users can see the top risky words and their contributions to the document’s total risk.
• —Risk Levels: The app classifies the document’s overall risk as Low (0-3), Moderate (3-6), or High (6-10) based on cumulative scores.
• —Customizable: Easily update the CSV dataset to add new risky terms relevant to your domain.
• —Efficient Processing: Uses AI embeddings to handle words not present in the dataset, while overfitting ensures scores are high only for dataset words.

This project is ideal for legal tech, contract analysis, compliance checks, and risk management applications. It demonstrates the integration of custom datasets, AI embeddings, and interactive visualizations in a modern web app.