Identifying Patterns in Kawasaki Disease Patients Through Statistical Analysis and Machine Learning

September 2020 - April 2021

Summary: Applied statistical analysis, data visualization, and machine learning to analyze a dataset of 1,000+ Kawasaki disease patients, including demographics, treatment response, and gene data. Used Z-tests, standard deviation analysis, decision trees, and random forests to identify relationships among clinical and genetic factors affecting IVIG responsiveness and disease outcomes. Found that ~98% of patients carried the GG allele in KCNN2 and 74% carried the TT allele in CRP, with nonresponders to IVIG showing significantly higher risk of coronary artery aneurysms, providing insights toward improved understanding and earlier diagnosis.

Programming language: Python