Zestimate — Housing Price Prediction
Machine learning model predicting home values using Zillow-like datasets with advanced feature engineering.
Overview
This project focused on predicting home sale prices using a large real-estate dataset inspired by Zillow. By combining exploratory data analysis, feature engineering, and regression models, I built a predictive pipeline that significantly improved on baseline estimates.
Approach
- Exploratory Data Analysis to identify key housing features influencing price.
- Feature Engineering to capture location, square footage, and derived metrics.
- Random Forest Regressor for robust non-linear modeling.
- Preprocessing Pipelines automated with Pandas & Scikit-learn.
- Hyperparameter Tuning for model accuracy.
Results
- Reduced margin of error compared to baseline pricing models.
- Automated preprocessing pipeline improved reproducibility and training efficiency.
Skills demonstrated: regression modeling, feature engineering, pipeline automation, hyperparameter tuning.