Financial Report Summarizer 📈

A comprehensive tool for analyzing and summarizing SEC 10-K filings using TF-IDF text summarization techniques. This application extracts, processes, and visualizes financial data from major tech companies (AAPL, GOOG, NVDA, TSLA etc.) directly from the SEC EDGAR database.

Homepage

Summary

Graphs

Features

• Automated SEC Data Scraping: Fetch 10-K filings directly from SEC EDGAR database
• Intelligent Text Summarization: TF-IDF based summarization with customizable compression levels
• Interactive Web Interface: Streamlit-powered UI for easy interaction
• Financial Data Visualization: Dynamic charts showing key financial metrics over time
• Multi-Section Analysis: Process different sections of 10-K filings (Part I, Item 7A, Item 9A)
• Data Processing Pipeline: Complete workflow from raw HTML to structured analysis

Installation

1. Clone the repository

git clone https://github.com/omkarsoak/Financial-Report-Summarizer.git
cd Financial-Report-Summarizer

2. Install required packages

pip install streamlit pandas matplotlib nltk requests beautifulsoup4 html2text

3. Download NLTK data (first-time setup)

import nltk
nltk.download('punkt')
nltk.download('stopwords')

Usage

Running the Web Application

streamlit run src/app.py

The application will start on http://localhost:8501

Application Workflow

1. Select Company: Choose from AAPL, GOOG, NVDA, or TSLA
2. Set Summarization Level:
   • 0.1-0.9: Large summary (more detailed)
   • 1.0: Medium summary (balanced)
   • 1.1-2.0: Small summary (highly compressed)
3. Generate Summaries: Automatic processing of different 10-K sections
4. View Financial Charts: Interactive visualizations of key metrics
5. Navigate: Use built-in navigation to move between sections

Data Processing Pipeline

The application follows this data flow:

SEC EDGAR → HTML Files → Text Extraction → TF-IDF Processing → Summary Generation → Visualization

Simplified version

A simplified version of the app is available at src/tf_idf_summarizer_simplified.py. To run:

python tf_idf_summarizer_simplified.py --ticker NVDA --n 1.5

The output is generated in the directory ./tf-idf-summary