Incorporated databased consumer-behavior models into research data processing with pySpark and spark.sql libraries. Performed Linear Regressions on selected permutations of attributes and proposed 6 product-prototyping strategies.

Helped refine the consumption-pattern-recognition model by removing 130 strongly-correlated hyperparameters with `dropouts' operation, resulting in an average 4\% increase in model accuracy.

Assisted in comparative analysis of key features that predict parental purchase decisions.

Incorporated Alibaba-powered database DMP to refine research on online purchasing behavior.