Yawei Han

Offering expertise in data analysis, forecasting, and business intelligence to enhance departmental effectiveness and support strategic decision-making. Skilled in Python, SQL, and R for data modeling and benchmarking, with proficiency in Tableau to deliver clear, actionable insights. Experienced in identifying trends, analyzing performance metrics, and implementing process improvements to drive operational efficiency. Recognized for strong analytical skills, effective communication, and the ability to translate complex data into business terms.

Hidden Markov Model Applied in Automatic Speech Recognition

Honour Thesis Sep. 2024 - Present

• Developed and implemented a Hidden Markov Model (HMM) for automatic speech recognition using Python, including a
• Extracted and analyzed acoustic features from speech signals using Librosa, leveraging MFCC features to enhance model accuracy and performance.
• Conducted preliminary analysis and visualization of audio feature data using R, and utilized HMM for building and training acoustic models, gaining hands-on experience in the entire process from data processing to model training and performance evaluation.

Dashboard Optimization and Student Behavior Analysis for Online Education Platform 

 Aug. 2024

• Conducted comprehensive analysis for user engagement and course performance for an educational platform, using SQL to extract data from 12 databases to assess user engagement, conversion rates, and the impact of marketing campaigns.
• Developed a strategic 5-page Tableau dashboard featuring KPIs, retention curves, and funnel visualizations to support product and marketing teams, confirming a significant increase in onboarding rate due to marketing campaigns and new feature launches.
• Analyzed user behavior data of over 19K students and proposed enhancements in user experience, marketing strategies, and content optimization to boost engagement and conversion rates.

Impact of Correlated Predictors on Machine Learning Performance

 Jan. 2023 - Apr. 2023

• Evaluated and compared the performance of machine learning models in the presence of high correlation and collinearity among predictors using regularization techniques (Linear, Lasso, Ridge, Elastic Net) and ensemble methods (decision tree, random forest).
• Simulated datasets with varying levels of correlation among predictors using R Studio and conducted statistical analyses to assess the impact on model performance, achieving a significant reduction in Mean Squared Error (MSE) across different correlation levels.
• Utilized R Studio for dataset simulation, statistical analysis, and result visualization, demonstrating effective mitigation of correlated predictors' impact and improved prediction accuracy with Lasso and Ridge methods.

Data Analysis of Exercise Effects on Heart Rate and Calorie Burn 

 Sep. 2022 – Nov. 2022

• Led a 2*2 factorial experiment to analyze the relationship between exercise intensity, duration, heart rate, and calories burned, using regression analysis, ANOVA, and various plots (box, residual, interaction, main effects, contour, 3D).
• Conducted comprehensive statistical analysis, revealing that heart rate is influenced by speed but not by workout duration or interaction effects, while calories burned are significantly affected by both speed and time.
• Utilized advanced statistical tools and methods in R Studio to generate detailed visualizations and reports, providing actionable insights for designing effective cardio exercise programs.