Stock Market Prediction System Final Year Project

• Secure admin login in this final year project
• Admin role-based access
• Separate admin console
• Violet/indigo admin theme
• Admin dashboard
• View total users
• View active users
• View inactive users
• View total announcements
• View active featured stocks
• View aggregate watchlist items
• View recent registrations
• Quick links to users, announcements, and featured stocks
• User management
• List all users
• Search users by name
• Search users by email
• Filter users by role
• Filter users by active/inactive status
• Create user
• Add name, email, password, role, and status
• Edit user
• Update name
• Update email
• Update role
• Update status
• Reset user password
• Delete user
• Prevent deletion of own admin account
• Assign user role
• Promote user to admin
• Change admin to user where allowed
• Deactivate user account
• Block login without deleting account
• Announcement management
• Create platform announcement
• Add title and message
• Select announcement type: info, success, warning, critical
• Set announcement priority
• Set announcement active/inactive status
• Edit announcement
• Delete announcement
• Toggle announcement visibility on public landing page
• Featured stock management
• Add featured stock
• Add stock symbol
• Add company name
• Add sector
• Add description
• Control display order on landing page
• Toggle featured stock active/inactive status
• Edit featured stock
• Delete featured stock
• Link featured stock cards to analyzer
• Manage public landing page content through announcements and featured stocks

The admin panel includes dashboard, user management, announcements, and featured stocks modules with platform-level CRUD controls.

• User registration in this final year project
• User login
• JWT token stored in localStorage
• Redirect after login
• Error handling for invalid credentials
• Deactivated account login blocking
• Auto-login after successful registration
• Public stock analyzer access
• Search Indian NSE stocks
• Search by company name
• Search by stock symbol
• Select custom start date
• Select custom end date
• Fetch OHLCV historical data
• View Open, High, and Close price chart
• View historical stock data table
• Highlight minimum and maximum values
• Export stock data to Excel
• Save stock to watchlist after login
• Run LSTM prediction
• View AI prediction page
• Generate 30-day LSTM forecast
• View combined historical and predicted chart
• View educational-use disclaimer
• View financial news feed
• Browse news cards
• View title, author, date, image, and external link
• Use paginated news feed
• Daily local news cache
• User dashboard
• View watchlist count
• View available stocks
• View account status
• Add stock to watchlist
• Add notes to watchlist item
• View watchlist table
• View saved stock symbol and name
• Edit watchlist notes
• Remove stock from watchlist
• Quick analyze saved stocks
• Profile management
• View own profile
• Update name
• Update email
• Change password
• Sign out
• Delete account permanently

The user-side modules include stock analyzer, AI prediction, market news, login, register, dashboard, watchlist CRUD, and profile CRUD.

• MERN-style full-stack architecture in this final year project
• Python ML pipeline integration
• React 18 frontend
• React Router 6 routing
• Tailwind CSS fintech-style interface
• Chart.js 4 interactive charts
• React Toastify notifications
• React Icons
• XLSX Excel export
• Node.js backend
• Express 4 REST API
• MongoDB database
• Mongoose 8 ODM
• JWT authentication
• bcryptjs password hashing
• CORS support
• Python 3.10+ scripts
• TensorFlow/Keras LSTM model
• yfinance data fetch
• pandas data processing
• scikit-learn MinMaxScaler and metrics
• NumPy 1.x compatibility for TensorFlow
• Financial Modeling Prep news support
• Alpha Vantage fallback support
• Yahoo Finance fallback support
• Multi-provider stock data fetch
• 24-hour memory and disk cache for stock data
• Admin-curated landing content
• Inactive user blocking
• Role-based admin access
• Protected routes
• Excel export of stock table
• Watchlist CRUD
• Announcements CRUD
• Featured stocks CRUD
• Educational and research use disclaimer

The technology stack includes React, Tailwind CSS, Chart.js, Node.js, Express, MongoDB, Mongoose, JWT, Python, TensorFlow/Keras, yfinance, pandas, scikit-learn, NumPy, and optional financial data APIs.

Technology Used

Data Sources

The project uses a multi-provider fallback chain for stock data:

1. Python yfinance as the primary source for NSE .NS tickers
2. Alpha Vantage .BSE fallback when API key is configured
3. Financial Modeling Prep when API key is configured
4. Yahoo Finance as a last-resort fallback through yahoo-finance2

Stock data results are cached for 24 hours, and LSTM predictions use get_stockdata.py with TensorFlow/Keras on historical close prices.

Step 1: Open the Project

cd stockMarketPrediction

Step 2: Configure backend environment

cd backend

copy .env.example .env

For Mac/Linux:

cp .env.example .env

Step 3: Add backend environment variables

Edit backend/.env:

MONGO_URI=mongodb://127.0.0.1:27017/stockvision

JWT_SECRET=your_strong_jwt_secret

JWT_EXPIRES_IN=7d

CLIENT_URL=http://localhost:3000

PORT=3001

[email protected]

ADMIN_PASSWORD=Admin@123456

ADMIN_NAME=Platform Admin

ALPHA_VANTAGE_API_KEY=

FINANCIALMODELAPI=

Backend environment variables include MongoDB URI, JWT secret, token expiry, frontend URL, backend port, admin seed credentials, and optional stock/news API keys.

Step 4: Start MongoDB

For Windows:

net start MongoDB

For Mac:

brew services start mongodb-community

For Linux:

sudo systemctl start mongod

Or use MongoDB Atlas and set MONGO_URI in .env.

Step 5: Install and run backend

cd backend

npm install

node scripts/seedAdmin.js

node index.js

Expected output:

Server started on port 3001

MongoDB connected

Step 6: Install Python ML dependencies

cd backend

pip install -r requirements.txt

Verify installation:

python -c "import tensorflow; import yfinance; print('Python ML OK')"

NumPy must be below version 2.0 for TensorFlow compatibility, and the requirements file pins numpy>=1.26.0,<2.0.0.

Step 7: Install and run frontend

Open a new terminal:

cd frontend

npm install

npm start

Frontend opens at:

http://localhost:3000

Backend API runs at:

http://localhost:3001

The local ports are frontend 3000, backend API 3001, and MongoDB 27017

Admin Account

Created by running:

node scripts/seedAdmin.js

Admin seed credentials can also be customized using ADMIN_EMAIL, ADMIN_PASSWORD, and ADMIN_NAME in backend/.env.

Regular User Account

No default regular user exists. Create one using:

• Register page: http://localhost:3000/register
• Admin panel: Admin → Users → Add User