Developed and validated a Random Forest Regression model to predict median home values in California. The project achieved an R² score of 0.803, demonstrating strong predictive accuracy and delivering clean outputs ready for Business Intelligence integration.

Real estate valuation requires accurate predictions to support investment decisions and BI reporting. The challenge was to build a robust regression model that could handle geographic and socioeconomic variables while producing interpretable outputs for stakeholders.

Data Acquisition & EDA: Loaded the California housing dataset, analyzed correlations, and visualized geographic price distributions.

Feature Engineering: Applied Train/Test split and standardized predictor variables.

Modeling: Implemented a Random Forest Regressor, tuned parameters, and validated performance.

BI Integration: Exported predictions into a clean CSV format for Power BI dashboards.

Delivered a high‑performance regression pipeline that combines statistical rigor with business usability. The model outputs were visualized in Power BI, enabling clear interpretation of prediction accuracy and error distribution.

Accuracy: Explained 80.3% of housing price variance.

Interpretability: BI dashboards provided geographic error maps, prediction vs. reality curves, and error histograms.

Business Value: Enabled stakeholders to make data‑driven decisions with confidence in model reliability.