shiny

This repository contains the code for R Shiny apps I have created, mostly for the purpose of teaching and learning statistics. Currently, the web apps are only accessible within the GVSU network, so sharing the code is the only way I have of sharing them outside the university. I have organized them as follows.

steps

This is a tool for starting to learn coding your own apps. It contains step-by-step instructions for making a simple app that plots the normal density curve. The instructions are an R Markdown file that contains several Shiny apps (code and web apps) at the different stages of development.

introstats

This folder contains several apps that are designed for an introductory course in applied statistics. These apps include:

• samp_dist_samp_prop Sampling distribution of the sample proportion
• ci_prop Confidence interval for a population proportion
• ht_prop Hypothesis test for a population proportion
• quant_descr Numerical and graphical summaries for the distribution of a quantitative variable
• guess_sd Given four histograms/boxplots and four standard deviation values, can you match the distribution to the standard deviation? (made by Suchir Gupta)
• clt The sampling distribution of the sample mean and the Central Limit Theorem
• std_norm "Forward" and "Backward" calculations for the standard normal distribution
• t_dist Comparing the t distribution to the standard normal distribution (and t* to z*)
• ci_mean Confidence intervals for a population mean
• ht_mean Hypothesis tests for a population mean
• chisq_test Chi-squared test for a two-way table
• slope_intercept Slope-intercept form of a line
• least_squares The least squares regression line has the smallest sum of squared residuals out of all lines.
• two_sample_apps Simulating independent two-group data; confidence intervals and hypothesis tests for a difference in population means
• testing_errors Type 1 and 2 errors and power illustrated through simulation in the two-sample t test context

regression_doe

This folder contains apps I have used in my courses in Regression and Design of Experiments, both at the undergraduate and graduate levels. The apps include:

• slr_model Simulate from the simple linear regression model. The focus is on the difference between the parameters and their estimates and on the differences between the errors and the residuals.
• power_curve Power curve for the two-sample t test, focusing on its dependence on the effect size, standard deviation, sample size, and significance level
• norm_quant_plot Produces normal quantile plots for simulated data from different distributions (normal and non-normal) and explains what is plotted on the x- and y-axes
• leverage Displays how leverage is a measure of the statistical distance of the x-values for an individual from the center of the x-values for the dataset, focusing on the case of two explanatory variables
• model_sel Shows the bias/variance tradeoff involved in model selection. Specifically, underfitting can lead to bias in regression coefficient estiamtes and overfitting can lead to increased variance.
• 1wayrandom Focuses on the difference between the statistical properties of fixed and random effects in one-way models
• 3wayANOVA Shows the effects of each term in the three-way factorial model on the cell means
• sim2factor Uses simulation to explore the differences between two-factor models with crossed/nested and fixed/mixed/random factors

misc

Contains Shiny apps about other subjects, including:

• spelling_app Can you choose between correctly and incorrectly spelled versions of 21 difficult words?
• trig_unit_circle Geometric interpretation of the sine and cosine functions using a circle with radius of one