House parice prediction with ML

Project Description:

Built an end-to-end machine learning pipeline to predict residential housing prices using structured real estate data. The project focused on transforming raw property data into predictive spatial and structural features while applying industry-standard regression modeling practices.
Project Objective: To clean, standardize, and impute missing data from a real estate dataset with nearly 10,000 records, ensuring the data is accurate, complete, and usable for downstream analysis or predictive modeling.

The workflow included:

  • Data cleaning and preprocessing
  • Missing Value analysis
  • Exploratory data analysis
  • Spatial feature engineering
  • Feature interaction creation
  • XGBoost regression modeling
  • Model evaluation in both log-space and real-dollar values

Outcome

The final model successfully captured pricing behavior across geographic regions while maintaining interpretable and production-conscious design principles.

-----------------------------------------------------------------------------------------------------------------

Music Clustering and Acoustic Similarity Analysis

Project Description:

Built a machine learning workflow to analyze and cluster songs based on acoustic characteristics and music metadata. The project focused on understanding audio similarity patterns and exploring how clustering techniques can support music recommendation systems.

The workflow included:

  • Audio feature preprocessing
  • Exploratory analysis of music attributes
  • Handling missing and low-quality records
  • Clustering model experimentation
  • Cluster quality evaluation using silhouette scoring
  • Recommendation-oriented similarity analysis

Outcome

The project demonstrates understanding of how unsupervised learning can support modern recommendation systems and music discovery platforms.