# This code will load the R packages we will use suppressPackageStartupMessages({ library(mosaic) library(supernova) }) font_size = function (size) { theme(text = element_text(size = size))} # Adjust scientific notation options(scipen = 999) middle <- function(x, prop = .95) { sorted <- sort(x) tail_size <- (1 - prop) / 2 upper_cut <- sorted[floor(length(x) * (1 - tail_size))] lower_cut <- sorted[ceiling(length(x) * tail_size)] ((x <= upper_cut) + (x > lower_cut)) == 2 } # Adjust the plots to have nice sizes options(repr.plot.width = 6, repr.plot.height = 4)

YouTubers <- read.csv("https://docs.google.com/spreadsheets/d/e/2PACX-1vRc-xoYTbcG_kCs0oNcR0uDgXn0MWroDw9e-EqLJeFDsRtNqXZJqt7A43IvhrSyjR8uEozsKjIy9EH3/pub?gid=615413346&single=true&output=csv", header=TRUE) head(YouTubers)

gf_histogram(~Income, data = YouTubers, color = "darkgreen", fill = "lightgreen")%>% gf_boxplot(color="darkblue", fill ="lightblue")

sdob <- do(3) * b0(Income ~ NULL, data = resample(YouTubers))

# to help us talk about "likely" samples in a sampling distribution # try running this code gf_histogram(~ b0, data = sdob, bins = 100, fill = ~middle(b0, .95))

# fit and save the empty model emptymodel <- lm(Income ~ NULL, data = YouTubers) # calculate the confidence interval confint(emptymodel)

# Compare the sample and the sampling distributions gf_dhistogram(~Income, data = YouTubers, color = "darkgreen", fill = "lightgreen", bins = 100) %>% gf_dhistogram(~b0, data = sdob, color = "gray", fill = "black", bins = 100)