Pengyu Zhao

Data Scientist @Idxexchange

CA, US

Developed a machine learning model to predict California property close prices, conducting data exploration across 100+ features and identifying top 10 key predictors through correlation analysis and visualization tools.• Developed an outlier detection model to identify abnormal close price via comparing list price and close price.• Compared multiple algorithms, including XGBoost and random forests, with XGBoost yielding the best results, improving model performance by 15% in RMSE.• Conducted hyperparameter tuning that resulted in a 20% performance improvement across evaluation metrics such as R-squared, MSE, and RMSE, ensuring accuracy and reliability in property price predictions.

Northwestern University

Master's degree, Data science

University of Wisconsin-Madison

Bachelor of Science - BS, Economics