🏠 House Price (House Rate) Predictor

A complete end-to-end Machine Learning project that predicts house/property prices based on structured data such as location, area, rooms, amenities, distance to city, and other real‑world features.

This project is built for learning + production readiness, covering:

• Data cleaning
• Feature engineering
• Model training
• Evaluation
• Inference
• Web deployment (Streamlit)

⚠️ Educational & research project. Predictions are estimates, not financial or legal advice.

To build an AI system that can:

• Learn patterns from historical property data
• Predict house prices accurately
• Explain predictions using model reasoning
• Provide a usable web interface for real users

• Machine Learning Type: Supervised Learning
• Task Type: Regression
• Input: Property features (area, rooms, location, etc.)
• Output: Predicted price (₹)

House-Rate-Predictor/
│
├── data/
│   ├── raw/                # original raw dataset
│   ├── interim/            # cleaned/intermediate data
│   └── processed/          # final training-ready data
│
├── notebooks/
│   └── 01_eda.py           # exploratory data analysis
│
├── src/
│   ├── config.py           # project config
│   ├── features.py         # feature engineering
│   ├── eda.py              # EDA utilities
│   ├── train.py            # model training
│   └── infer.py            # prediction script
│
├── app/
│   └── streamlit_app.py    # web application
│
├── models/                 # trained models
├── requirements.txt
└── README.md

• area_sqft
• bedrooms
• bathrooms
• age_yrs
• floor
• total_floors
• amenities_count
• distance_cbd_km
• distance_metro_km
• latitude
• longitude

• bath_per_bed = bathrooms / bedrooms
• floor_ratio = floor / total_floors
• amenity_density = amenities / area

• city
• furnishing
• builder
• property_type

• Load CSV dataset
• Validate schema
• Type conversion

• Missing value detection
• Outlier detection
• Distribution analysis
• Correlation analysis

• Missing value imputation
• Outlier winsorization
• Data normalization

• Ratio features
• Density features
• Robust transformations

• Algorithm: XGBoost Regressor
• Target transformation: log(price)
• Pipeline-based training
• Train/Validation split

• RMSE (Root Mean Squared Error)
• MAPE (Mean Absolute Percentage Error)

• Model saved using joblib
• Web UI using Streamlit
• Real-time predictions

• Model: XGBoost Regressor

• Loss Function: Mean Squared Error

• Target Transformation: log(price)

• Why log(price)?

  • Reduces skew
  • Stabilizes training
  • Improves prediction reliability

Built using Streamlit, allows users to:

• Input property details
• Upload features
• Predict price
• See prediction confidence band

# create virtual env
python -m venv .venv

# activate
.venv\Scripts\activate   # windows
source .venv/bin/activate # mac/linux

# install dependencies
pip install -r requirements.txt

python -m src.train

streamlit run app/streamlit_app.py

id,city,lat,lon,area_sqft,bedrooms,bathrooms,age_yrs,floor,total_floors,
amenities_count,distance_cbd_km,distance_metro_km,furnishing,builder,property_type,price,listed_date

• No personal data stored
• No identity tracking
• No financial decision automation
• Predictions are estimates only

• CatBoost model integration
• SHAP explainability
• Satellite image integration
• NLP-based property description parsing
• Location clustering
• RAG-based property advisory
• API deployment (FastAPI)
• Mobile app version

This project teaches:

• Real-world ML system design
• Production ML pipelines
• Data engineering concepts
• Model deployment
• AI system architecture
• End-to-end AI product building

Built as part of an AI learning journey to master:

• Machine Learning
• Generative AI
• AI Engineering
• Production AI systems

MIT License — Free to use, modify, and learn from.

Give the project a ⭐ and use it to build your AI portfolio.

This project is not just a model — it is a full AI system blueprint.