Email Spam Detection System Final Year Project with Source Code

• Separate admin login in this final year project
• Admin-only access to /admin/*
• Admin dashboard
• View total user count
• View global prediction statistics
• View recent users
• View recent predictions
• View latest model metrics snapshot
• View confusion matrix if model is trained
• User management
• List non-admin users
• Search users by name
• Search users by email
• View all predictions
• Search global prediction log
• Filter global predictions
• Delete inappropriate prediction rows
• Dataset management
• Upload CSV dataset to data/
• Select CSV file for training
• Train machine learning model
• Update dataset metadata
• View model metrics
• View accuracy score
• View precision score
• View recall score
• View F1 score
• View confusion matrix values
• View confusion matrix plot
• View training history log
• Train/test split support
• Best model selection by F1 score
• Saved model artifacts management through training pipeline

• User sign up in this final year project
• User login and logout
• Forgot password
• Reset password using registered email verification
• Secure password hashing using Werkzeug
• Session-based user access
• User dashboard
• View total emails checked
• View spam count
• View ham count
• View spam percentage
• View recent predictions
• Analyze email subject and body
• Combine subject and body for prediction
• Spam / Ham prediction result
• Prediction confidence score
• Prediction timestamp
• Save each prediction in history
• View prediction history
• Search history by keyword
• Filter records by spam
• Filter records by ham
• Filter all records
• Filter history by date range
• Delete individual prediction records
• View per-user statistics
• Dataset information page
• View last registered dataset stats
• Profile page
• Update name and email
• Change password with current password verification
• Export prediction history as CSV
• Export prediction history as plain text
• Normal user routes protected under /app/*
• Admin users redirected to admin console

Machine Learning Features in this Final Year Project

• Email spam/ham classification
• Locally trained machine learning model
• No third-party inference API
• TF-IDF vectorization
• Naive Bayes model training
• Logistic Regression model training
• Linear SVM model training
• Best model saved by F1 score
• Train/test split
• Prediction preprocessing aligned with training preprocessing
• NLTK stopword handling
• NLTK tokenization
• NLTK lemmatization
• pandas CSV data loading
• joblib model persistence
• matplotlib confusion matrix plot
• Dataset metadata tracking
• Training metrics storage
• Requires trained model before normal users can analyze emails

• Model artifacts saved under:

  model_artifacts/

Other Features in this Final Year Project

• Complete email spam detection source code
• Suitable for final year project, major project, and minor project
• Project report content can be prepared from included modules
• Flask web application
• SQLite database
• Flask-SQLAlchemy ORM
• User authentication
• Admin authentication
• Secure password hashing
• Spam/ham prediction history
• User statistics
• Admin dataset upload
• Admin model training
• Model metrics display
• Confusion matrix plot
• CSV export support
• Text export support
• Kaggle dataset downloader
• Sample dataset included
• No external inference API required
• Useful for viva, source code review, project report writing, and project demonstration

1. Open the project folder that contains:

   run.py
   requirements.txt

2. Create virtual environment:

   python -m venv .venv

3. Activate virtual environment on Windows:

   .\.venv\Scripts\Activate.ps1

4. Install dependencies:

   python -m pip install -r requirements.txt

5. Seed the database:

   python seed.py

6. Optional: download Kaggle dataset:

   python download_kaggle_dataset.py

7. Start the application:

   python run.py

8. Open project in browser:

   http://127.0.0.1:5000

9. Open user login:

   /auth/login

10. Open admin login:

/auth/admin_login

How to train the model?

1. Start the Flask app:

   python run.py

2. Open admin login:

   http://127.0.0.1:5000/auth/admin_login

3. Login using admin credentials.

4. Open Dataset & Training:

   http://127.0.0.1:5000/admin/dataset

5. Select sample dataset:

   sample_sms_demo.csv

6. Click:

   Train & evaluate

7. After training, open Model Metrics to view accuracy, precision, recall, F1 score, and confusion matrix.
8. Normal users can then use Analyze Email.

Administrator Account

• Email: [email protected]
• Password: Adminpass1
• Admin Login: /auth/admin_login

Demo User Accounts
All demo users use password:

Userpass1

Demo users:

• [email protected]
• [email protected]
• [email protected]
• [email protected]
• [email protected]
• [email protected]
• [email protected]

User Login:
/auth/login