In Tom-bomb/glowup: Tom's assigned R package (python is better)

knitr::opts_chunk$set(
  fig.width = 10,
  fig.height = 6,
  collapse = TRUE,
  comment = "#>"
)

library(glowup)

Basics

To demonstrate my deep and intimate knowledge with all things R, We will take a look at a most useful glowup function. If you, dear reader, ever need the word derp printed out on demand, this package is just the thing for you. Observe:

print_derp()

Wow, that was really something right?

Let's Get Into it

We will now demonstrate a fundamental concept in stats - the Central Limit Theorem. Given enough observations, the sampling distribution will always be normal, despite whatever distribution the sample's population has.We will demonstrate on my myclt function, which will sample from a uniform population:

tmp = myclt(30, 10000)

Look at that! Sure looks normal to me.

Holy Stats Man

Next we will demonstrate the concept of bootstrapping. Let's say that a sample was really hard or costly to obtain, but the sample you did get with your grant money was a good representation of the population in question. We can use bootstrapping to extract a bit more out of these data by resampling the sample. Observe:

sam = rnorm(20, mean=0, sd=10)
myboot2(x=sam, alpha=0.005)

I am Confident in STATS

Let's finish up our package demonstration by doing a classic confidence interval estimation. This concept is wicked important for many reasons, but the reason I believe it is crucial is because the idea of a confidence interval is the basis for running statistical tests on data or sets of data. Let's go simple and look at getting a point estimate for the mean and estimate where the population's true mean is likely to truly reside:

sam = rnorm(2550, mean=0, sd=3)
myci(sam, alpha=0.5)

Going FURTHER

I have made a shiny APP! See .rbf file for demo.

Tom-bomb/glowup documentation built on Dec. 31, 2020, 5:32 p.m.