• Processed 100k data by conducting data cleaning and manipulation to improve the model 80% performance on classification models in R (tidyverse, janitor, broom) to predict the likelihood an applicant to be selected by USCIS H1B lottery system.
• Created H1B visa lottery algorithm that forecasts to approximately 80% actual values from 6 different machine learning classification methods in R (glmnet, MASS, tidymodels, pls, ISLR2, boot, class, caret), including K-Nearest Neighbor, Logistic Regression, Ridge/LASSO Regression, Partial Least Squares and Principal Component Regression.