How faizan-fayaz-10/house-price-prediction. works
Main Components
a. Data Source
• Boston Housing dataset (loaded via scikit-learn)
b. Data Processing & EDA
• Loading & cleaning raw data
• Generating exploratory visuals (correlation heatmap)
c. Model Training
• Linear Regression implementation (scikit-learn)
d. Evaluation & Visualization
• Calculating R², RMSE
• Plotting predicted vs. actual prices
e. Model Persistence
• Saving the trained model to house_price_model.pkl using joblib
f. Outputs
• Saved images (correlation_heatmap.png, predictions.png)
• Persisted model file
Relationships & Data Flow
• Data Source → Data Processing & EDA
• Processed Data → Model Training
• Trained Model + Test Data → Evaluation & Visualization
• Trained Model → Model Persistence
• EDA & Evaluation steps → Output files
Architectural Patterns & Principles
• Linear, pipeline-style flow—each stage feeds the next
• Separation of concerns: ingestion, analysis, training, evaluation, persistence
• Reproducibility—model persistence for deployment or later inference
Key Technologies & Frameworks
• Python 3
• scikit-learn (data loading + LinearRegression)
• pandas / numpy (data manipulation)
• matplotlib / seaborn (visualizations)
• joblib (model serialization)
Putting It All Together
[ Dataset (sklearn) ]
↓
[ Data Processing & EDA (pandas, seaborn) ]
↓
[ Model Training (LinearRegression) ]
↓
[ Evaluation & Visualization (metrics, matplotlib) ]
↓
[ Model Persistence (joblib) ]
↓
[ Output Files: .png visuals, .pkl model ]