Hang Yang

Experience

• 

  PepsiCo

  Jul 2020 - Jan 2021

  Supply Chain Management Analyst

  - Leveraged Power BI to analyze inventory turnover rates in multiple cities, leading to data-driven sales strategy enhancements- Conducted Web Scraping of PepsiCo's online store data to identify key factors influencing customer satisfaction. Created product satisfaction level categories using word cloud visualizations, guiding targeted product improvement efforts- Implemented Power BI incorporating historical sales data and market trends. Improved inventory planning and waste reduction, enhancing supply chain efficiency and aligning production with market demands Show less

• 

  Sina.com

  Jan 2021 - Mar 2021

  Data Analyst

  - Generated dozens of user portraits under phone branding topics based on User Behavior Analytics and presented the result to senior leadership, providing a comprehensive summary of marketing strategies and key competitors- Conducted Weibo data platform research by gathering useful data to develop an understanding of customer behavior, demographics, and life cycle of a local brand in China; presented data that helped guide decisions of that company, thereby raising its popularity 43%

• 

  Bilibili Group

  Apr 2021 - Jul 2021

  Data Analyst

  - Employe Power Query and Pandas package in Python to perform data cleaning, visualized the results by using matplotlib packages, and converted over 600,000+ entrie of data to be more readable, and user-friendly- Scraped data from 100+ competitor account by using BeautifulSoupand Selenium to monitor competitor behavior, conducted Python TF-IDF analysis generated Word Frequencies table and increased the competitor behavior prediction rate by 70%- Utilized Python to design a competitor monitoring system, which automatically scraped data from multiple competitor accounts(by day), and optimized the procedure by 90% Show less

• 

  Chewy

  Jun 2022 - Aug 2022

  Data Scientist

  - Utilized Google Analytics to retrieve and analyze user behavior data, contributed over 800 lines of SQL query in Snowflake; Build an ecommerce conversion funnel with Tableau, successfully pinpointing key drop-off points for in-depth analysis- Employed Tableau to ascertain customer category interest levels, identified the top categories for transactional customers, and developed 3 targeted marketing strategies based on data insights- Formulated targeted marketing strategies based on transactional customer top existing page, significantly addressing checkout page abandonment and enhancing order conversion rates.- Implemented an XGBoost model to identify new customers unlikely to place orders. Enabled pinpointing user characteristics linked to inactivity, early identification of 8K at-risk customers, informed subsequent targeted engagement strategies Show less

• 

  NYU Langone Health

  Sept 2022 - Nov 2022

  Data Scientist

  - Developed Ridge and Lasso Regression to predict variables akin to hospital stay durations, focusing on enhancing predictive accuracy while preventing overfitting. Demonstrated model efficiency with an R square of 83%- Created over 10 Tableau dashboards, translating published medical studies into visually intuitive charts effectively selecting relevant patient/study groups, enhancing decision-making processes based on data-driven insights- Conducted PCA to reduce the dimension from 44 original features to 5 principal components. Subsequently applied the K-means algorithm to segment a large dataset with over 70,000 entries into 4 clusters Show less

• 

  Chewy

  Jun 2023 - Nov 2023

  Data Scientist

  - Conducted Survival Analysis and Random Forest to categorize user profiles and predict customer tenure, successfully identifying the top three churn drivers and 13K potential churn customer in October- Developed and validated a business hypothesis through A/B testing, assessing impact on customer engagement strategies, successfully demonstrated a 12% decrease in customer churn rate- Optimized delivery delay forecasts using AWS SageMaker XGBoost with RFE feature selection, achieving 83% precision.- Implemented a priority shipping strategy for at-risk customers by using prediction model, contributing to an 8% improvement in retention rates- Utilized KNN clustering for accurate prediction of waiting times in fulfillment centers. Integrated with XGbregressor to develop an effective picking strategy, enhancing unit picking throughput by 2.7%, average time saving of 14s per order pick- Collaborated across teams to establish an initial data quality check pipeline with Great Expectations to support developing MLOPS Show less

• 

  Einride

  Apr 2024 - now

  Business Analyst

  - Solution Architecture