jung-pu-chang Goto Github PK

2022.9 ~ now - Data Analyst at Datarget

• Using GNN + SNA + Collaborative Filtering to build recommandation system and sale to famous retail enterprises. Recommandation system promote customer's demand and take huge benefit to famous retail enterprises.
• Using Binning + Feature Selection + Regression to build member rating system and sale to famous retail enterprises. From the system, we can figure out the most suitable member in every activities and explain the reason.
• Using Prophet + SARIMAX + Feature Creation to build Inventory Management System. POC in famous semiconductor manufacturing enterprise. It reduced 70% deviation.
• Using GA optimization to build retail map and sale to famous retail enterprises. Retail map can optimize teant place, bring confortable shopping experience to customer.

2022.4 ~ 2022.7 - Data Analyst Intern at Yulon Finance Corporation

• Using Feature Selection(IV、Fisher、GINI、AIC.etc) to figure out useful feature for default model. Finally, we promote 20% KS value and avoid many clients default.

2020.9 ~ 2022.6 - Research (Feature Engineering) Assistant & Master Student in DS.SCU

• Reducing Data Set Dimensionality by Fu-Jung Feature Selection Method - Using Business Application Problem as an Example

2018.8 ~ 2020.6 - Data Analyst at CashBox

• Using associate analysis to build recommandation system and import to KTV. After customer choose the song, system would recommand food or drink to customer.
• Using xgb to predict consumer spending power and caculate suitable coupon.

2014.9 ~ 2018.6 - Bachelor student in AM.NDHU