• Monitored KPIs on advertising cost and return in a daily basis throughout 28 channels regarding ROI, channel traffic, and user quality, decreasing the rate of defective advertising by 17%• Forecasted user conversion with Lasso Regression Model, raising the prediction accuracy by 34%• Optimized channel budget distribution via contribution analysis of retention, improving retention rate by 7% • Identified KPIs on channel performance from the perspective of users, buzz and engagement for evaluation• Proposed the stepwise rewards for sparkling channels, developing overall ROI by 12%

• Automated data pipeline by creating scripts in Hive, Spark SQL and Python, raising operational efficiency by 40%• Created automated dashboards on Tableau to monitor crime incidents in support of safety strategies• Proposed an authentication mechanism for suspicious users with required upload of credit card via A/B Testing, resulting in the decline rate of crime incidents within 3 months nearly ten times that of order volume