Project Details
This graduate capstone project is a testament to the power of statistical modeling in healthcare research. It focuses on applying Linear Mixed Models (LMMs) within R to scrutinize the effect of BMI on intraocular surgery (IOS) measures, leveraging a diverse dataset from Kaggle. The data’s complexity, reflecting variations in IOS outcomes attributable to BMI differences, provided a rich learning ground for employing advanced statistical methodologies.
Through detailed exploratory data analysis, we unraveled intricate patterns and discerned the subtle influences of BMI on IOS metrics, utilizing visual tools to elucidate these relationships. The modeling phase involved rigorous LMM fitting, with model selection guided by AIC and BIC considerations, culminating in a model that embodies random slopes and intercepts to capture the variability of IOS responses across individuals.
Our diagnostic evaluations confirmed the robustness of the LMM, ensuring the assumptions of homoscedasticity and normality were satisfied. The project’s findings reveal a notable correlation between BMI and IOS outcomes, effectively encompassed by the LMM’s random effects.
The academic journey of this capstone has been as informative as it has been transformative, advancing our comprehension of both the subject matter and the analytic tools at our disposal. The project highlights the indispensability of LMMs in handling complex, real-world data in medical contexts, reinforcing the relevance of data science in driving medical research innovation. As emerging data scientists, the hands-on experience and insights garnered from this study are invaluable, solidifying our foundation in statistical analysis and preparing us for future challenges in the field.
This project was completed under the guidance of Dr. Achraf Cohen (https://pages.uwf.edu/acohen/)
Project Highlights
✅ Course completed with an ‘A’
✅ Academic Group Capstone Project - Part of UWF M.S. Data Science (https://uwf.edu/programs/hmcse/data-science-ms/)
✅ Comprehensive Literature Review on LMMs
✅ Full By-Example on LMM Implementation with R
✅ Final Report, Code, and Presentation Publicly Available
