Visualized health insurance original data by time frame to find the trend of the selling record by using correlation matrix in R
Cleaned and split data with Tidymodels package to select best performed column
Built several models such as Boosted Tree Model, Logistic Regression Model and Random Forest Model to demonstrate test dataset and select best performed model.
Applied Logistic Regression Model to the test dataset by plotting confusion matrix and calculate corresponding accuracy and AUC value.
Time Series Project
01.2022 - 03.2022
Selected data from U.S Candy Production in 12-months range to detect re-occuring pattern at fixed frequency
Utilized log transformation technique to achieve stationary series with stable variance
Designed SARIMA model in R and fitted parameters for model to best demonstrate candy production in time frame
Forecasted data in 3-month interval and compared with original data to prove the validity of model.
Analytic Project
02.2021 - 05.2021
Cleaned and analyzed data from 2016 and 2020 presidential polls to demonstrate distribution of votes in several states by using R
Predicted and confirmed polls of two elections by generating scatter plots and graphs to compare each dataset within small-time frame
Orchestrated and manged detailed forecasts with R and Excel for items, SKU across more than 100 channels for upcoming weeks, months and years of demand in pieces, first cost and gross revenue as well as overseeing a cumulative monthly shipping and reordering plan across diverse collections. Introduced new products and phased out underperforming ones based on the forecast for portfolio enhancement.
Utilized promotional opportunities on all channels(Wayfair, Amazon, Target, Home Depot etc.), offering different discounts to SKUs based on their inventory health while keeping a considerable weighted margin with VLOOKUP, Pivot Table and Index Match function in Microsoft Excel.
Optimized assortment selection through new SKU setup, data-oriented analytical projects, and rigorous discontinuation process.
Conducted Ad hoc Analysis report, descriptive analysis and developed and maintained BI applications and US Promotion dashboard on weekly basis to keep track of and analyze product performance and adjusted promotion strategy to reduce slow/aged inventory while improving overall profit margins.