FinCrack Web Page

FinCrack - Crack Finance with AI

🧠 Inspiration

Banking services are overwhelming to navigate, especially when financial needs vary by user background. We wanted to create a personal financial assistant that demystifies stock trends and recommends bank products tailored to individual profiles, especially for beginners with limited knowledge of banking services and stock markets.

💡 What it does

💹 1. Financial Dashboard and Analysis

Interactive dashboard providing real-time stock data, key ratios (PE, PB, P/S, PEG), profitability metrics, and technical indicators with visual trend analysis.

📊 Include real-time stock prices, PE/PB/P-S ratios, ROA/ROE metrics
📈 Feature technical indicators (RSI, SMA, EMA, MACD) with year-long price graphs

🏦 2. Bank Recommendation System

Offer personalized bank suggestions based on both qualitative preferences and product needs, matching users with suitable financial institutions.

👤 Analyze user demographics and preferences
🔍 Evaluate banks based on customer-specific requirements

💳 3. Financial Service Recommendations

Intelligently ranks financial services (loans, credit cards, mortgages) based on individual user profiles and needs.

💰 Consider factors like income, age, employment status, debt, and credit score
🏆 Provide ranked suggestions tailored to each user

🤖 4. Financial Chatbot

AI-powered assistant that translates complex financial data into clear, actionable insights in plain English.

❓ Answer finance-related questions
🧠 Powered by Gemini AI for natural language understanding

📰 5. News Sentiment Analysis

Analyze financial news to extract market sentiment around companies or financial topics.

🔍 Use Brave News Search API to gather recent articles
📊 Leverage Google Gemini to classify sentiment (Bullish/Bearish) with confidence scores
📝 Summarize articles and extracts key insights

🛠️ How we built it

Frontend: Node.js, Nest.js, MongoDB, FastAPI
Backend: Python, MongoDB
Machine Learning: Google Gemini AI, Pandas, Matplotlib
Infrastructure & Deployment: AWS, Docker

We used Google Colab and Jupyter Notebook for rapid prototyping, Streamlit for building the user interface, and GitHub for version control and collaboration.

🚧 Challenges we ran into

Integrate Gemini with function-calling logic in a conversational interface.
Handle edge cases in ticker symbol resolution.
Manage compatibility issues upon deployment.

🏆 Accomplishments that we're proud of

Created a functional multi-purpose financial assistant under 24 hours.
Integrated natural language stock analysis via Gemini.
Built a dual-system recommendation engine that bridges both finance education and practical guidance.

📚 What we learned

Prompt engineering for financial AI agents.
Merging generative AI with rule-based recommendations.
Handling finance APIs, user input normalization, and visualization in real time.