Interactive analysis of nonparametric statistical methods including sign tests, Wilcoxon rank-sum and signed-rank tests, Kruskal-Wallis, Spearman correlation, and Monte-Carlo simulation. Features 8 interactive plotly visualizations and 3 live browser-based simulators for hands-on exploration. Northeastern University ALY 6015 Assignment 5.

ALY6015 Module 5 final project. Lasso-regularized logistic regression on 68,061 diabetic inpatients (Strack et al. 2014), with bootstrap CIs, decision curve analysis, subgroup audit, and a live in-browser risk calculator powered by Ridge coefficients fit in the document.

An interactive walkthrough of nine hypothesis tests using Chi-Square goodness-of-fit, tests of independence, one-way ANOVA, and two-way ANOVA. Covers blood-type distributions, airline performance, movie admissions, military rosters, sodium content, sales data, school spending, plant growth, MLB baseball wins by decade, and a designed crop-yield experiment. Built in R with ggplot2 and Plotly for interactive visualizations, plain-English explanations, and stakeholder recommendations. ALY 6015 — Module 4

ALY6015 Module 3 — Logistic Regression analysis using the ISLR College dataset to predict whether a university is private or public. Includes EDA, glm() model fitting, confusion matrices, ROC curve (AUC = 0.976), standardised coefficient plot, and predicted probability distribution. Built with R, ggplot2, and Plotly.

Initial statistical analysis of 68,061 diabetic patient encounters from 130 US hospitals. Uses chi-square tests, ANOVA, and logistic regression to identify readmission predictors. Interactive Plotly visualizations show age effects, A1C testing impact, and a key finding: patients with 3+ prior admissions have 25-30% readmission rates. ALY6015 Module 4 Report.

An interactive R Markdown report for ALY6015 (Intermediate Analytics) at Northeastern University. This analysis applies Ridge regression, LASSO regression, and backward stepwise selection to the ISLR College dataset to predict graduation rates across 777 American colleges. The report features interactive plotly charts (hover for details, click to filter, drag to zoom), searchable coefficient tables, color-coded variable selection results, and a complete reproducible R code appendix organized by section. Key findings: LASSO achieves the lowest test RMSE (≈9.62) while automatically reducing the model from 17 to 8 predictors, outperforming both Ridge (9.77) and stepwise selection (9.76). All code is fully reproducible with fixed random seeds.