Code associated with blog post: http://t-redactyl.github.io/blog/2015/09/representative-sampling.html

comparing_small_and_larger_samples.R

set.seed(567)

# Sample of 30 (29 from the Poisson distribution and an outlier of 260)
sample1 <- c(rpois(29, lambda = 220), 260)

# Sample of 10 (9 from the Poisson distribution and an outlier of 260)
sample2 <- c(rpois(9, lambda = 220), 260)

generating_2014_ts_data.R

# Simulate some data for 2014, with mean page views of 220 per day.
days <- seq(as.Date("2014/1/1"), as.Date("2014/12/31"), "days")
page.views <- rpois(365, lambda = 220)
views <- data.frame(days, page.views)

# Generate post Christmas sale peak period (January 1-10) with mean page views of 400 per day.
views$page.views[views$days >= "2014/01/01" & views$days <= "2014/01/10"] <- rpois(10, lambda = 400)

# Generate Easter peak period (April 1-21) with mean page views of 350 per day.
views$page.views[views$days >= "2014/04/01" & views$days <= "2014/04/21"] <- rpois(21, lambda = 350)

# Generate Black Friday and Christmas peak periods (November 28-December 31) mean page views of 500 per day.
views$page.views[views$days >= "2014/11/28" & views$days <= "2014/12/31"] <- rpois(34, lambda = 500)

plotting_2014_ts_data.R

require(ggplot2); require(knitr); require(magrittr)
opts_chunk$set(fig.path='/figure/',echo=FALSE, warning=FALSE, message=FALSE)

qplot(days, page.views, data = views, geom = "line") %>%
    add(geom_line(colour = "#4271AE")) %>%
    add(ggtitle('Website Page Views in 2014')) %>%
    add(xlab('')) %>%
    add(ylab('Daily page views')) %>%
    add(geom_hline(yintercept=220, linetype="dashed")) %>%
    add(theme_bw())