Finsum: Demystifying Company Finances with AI 📊🤖

🌟 Project Overview

Finsum is an innovative AI-powered platform designed to break down the complex world of financial statements and make company insights accessible to everyone.

The Challenge 📈

Financial documents are notoriously complex:

• Filled with technical jargon
• Laden with intricate calculations
• Intimidating for those without a finance background

Our Solution 🧠

Finsum leverages cutting-edge AI to transform dense financial data into clear, digestible insights that anyone can understand.

🚀 Key Features

1. Company Story Explorer

• Select a company and get an AI-generated narrative of its financial journey
• Interactive financial metric visualizations
• Contextual AI chat to answer your specific questions

2. Finance Basics Learner

• Explore financial terms and concepts
• Receive beginner-friendly explanations
• Learn through simple analogies and plain language

🛠 Technology Stack

Frontend

• Streamlit

Backend & Data Processing

• Python
• Pandas, NumPy
• Web Scraping Tools (Selenium, BeautifulSoup)
• PDF Processing (PyMuPDF)

AI & Machine Learning

• OpenAI SDK (via OpenRouter)
• Langchain
• Transformers
• Sentence-Transformers

Data Storage

• Pinecone (Vector Database)
• PostgreSQL
• Local File Storage

Visualization

• Plotly
• Matplotlib
• Seaborn

📦 Project Structure

finsum/
│
├── .streamlit/           # Streamlit configuration
├── data/                 # Processed company data
├── data_extraction/      # Scripts for raw data retrieval
├── data_preprocessing/   # Data cleaning and formatting scripts
├── pages/                # Streamlit application pages
├── pipeline/             # Intermediate data processing
├── prompts/              # LLM prompt templates
├── utils/                # Utility scripts
├── Welcome.py            # Main Streamlit entry point
└── requirements.txt      # Project dependencies

🔧 Setup and Installation

Prerequisites

• Python 3.8+
• Git
• Chrome Browser (for web scraping)

Installation Steps

1. Clone the repository:

   git clone https://github.com/yourusername/finsum.git
   cd finsum

2. Create a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

4. Set up API Keys:

   • OpenRouter API Key (for AI features)
   • Pinecone API Key (for vector database)

   Store these in .streamlit/secrets.toml or .env

5. Install ChromeDriver:

   • Download matching your Chrome version
   • Update path in data_extraction/extract_latest_news.py

🚀 Running the Application

streamlit run Welcome.py