samplingsimulatorr

samplingsimulatorr is an R package intended to assist those teaching or learning basic statistical inference.

Authors

Name	GitHub
Holly Williams	hwilliams10
Lise Braaten	lisebraaten
Tao Guo	tguo9
Yue (Alex) Jiang	YueJiangMDSV

Overview

This package allows users to generate virtual populations which can be sampled from in order to compare and contrast sample vs sampling distributions for different sample sizes. The package also allows users to sample from the generated virtual population (or any other population), plot the distributions, and view summaries for the parameters of interest.

Installation

You can install the development version of samplingsimulatorr from GitHub with:

# install.packages("devtools")
devtools::install_github("UBC-MDS/samplingsimulatorr")

Function Descriptions

• generate_virtual_pop creates a virtual population.
  • Inputs : distribution function (i.e. rnorm, rexp, etc), the parameters required by the distribution function, and the size of the population.
  • Outputs: the virtual population as a tibble
• draw_samples generates samples of different sizes
  • Inputs : population to sample from, the sample size, and the number of samples
  • Outputs: returns a tibble with the sample number in one column and value in a second column.
• plot_sample_hist creates sample distributions for different sample sizes.
  • Inputs : population to sample from, the samples to plot, and a vector of the sample sizes
  • Outputs: returns a grid of sample distribution plots
• plot_sampling_dist creates sampling distributions for different sample sizes.
  • Inputs : samples created by draw_samples function, variable of interest, a vector of the sample sizes, and the number of replication for each sample size
  • Outputs: returns a list of sampling distribution plots
• stat_summary: returns a summary of the statistical parameters of interest
  • Inputs: population, samples, parameter(s) of interest
  • Outputs: summary tibble

How do these fit into the R ecosystem?

To the best of our knowledge, there is currently no existing R package with the specific functionality to create virtual populations and make the specific sample and sampling distributions described above. We do make use of many existing R packages and expand on them to make very specific functions. These include: - built-in r distribution functions such as rnorm to sample from distributions - rep_sample_n to generate random samples - ggplot2 to create plots Python pandas already includes some summary statistics functions such as .describe(), however our package will be more customizable. Our summary will only include the statistical parameters of interest and will provide a comparison between the sample, sampling, and true population parameters.

Dependencies

• dplyr >= 0.8.5
• rlang >= 0.4.5
• infer >= 0.5.1
• magrittr >= 1.5.0
• gridExtra >= 2.3.0
• ggplot2 >= 3.3.0

Usage

generate_virtual_pop

library(samplingsimulatorr)
generate_virtual_pop(N, var_name, dist, ... )

Arguments:

• N: The number of samples
• var_name: The variable name that we need to create
• dist: The distribution that we are generating samples from
• ...: The arguments required for the distribution function

Example:

pop <- generate_virtual_pop(100, "height", rnorm, 0, 1)

draw_samples

library(samplingsimulatorr)
draw_samples(pop, reps, sample_size)

Arguments:

• pop the virtual population as a tibble
• reps the number of replication for each sample size as an integer value
• sample_size the sample size for each one of the samples as an array

Example:

samples <- draw_samples(pop, 3, c(1, 10))

plot_sample_hist

library(samplingsimulatorr)
plot_sample_hist(pop, samples, var_name, sample_size)

Arguments:

• pop the virtual population as a tibble
• samples the samples as a tibble
• var_name the name of the column for the variable that is being generated
• sample_size a vector of the sample sizes (each sample size needs to be in the samples df input)

Example:

plot_sample_hist(pop, samples, height, c(1, 10))

plot_sampling_hist

library(samplingsimulatorr)
plot_sampling_hist(samples, var_name, sample_size)

Arguments:

• samples the samples as a tibble
• var_name the name of the column for the variable that is being generated
• sample_size a vector of the sample sizes (each sample size needs to be in the samples df input)

Example:

plot_sampling_hist(samples, height, c(10, 50), 100)

stat_summary

library(samplingsimulatorr)
stat_summary(pop, samples, paramater)

Arguments

• population the virtual population
• samples the drawn samples
• parameter the parameter(s) of interest

Example

stat_summary(pop, samples, c(mean, median))

Example Usage Scenario

library(samplingsimulatorr)

# generate population
pop <- generate_virtual_pop(1000, "height", rnorm, 0, 1)
head(pop)
#> # A tibble: 6 x 1
#>   height
#>    <dbl>
#> 1  1.07 
#> 2  0.707
#> 3  0.853
#> 4 -1.63 
#> 5 -0.512
#> 6 -0.833

# create samples
samples <- draw_samples(pop, 100, c(1, 10, 50, 100))
head(samples)
#> # A tibble: 6 x 4
#> # Groups:   replicate [6]
#>   replicate    height  size rep_size
#>       <int>     <dbl> <dbl>    <dbl>
#> 1         1  0.804        1      100
#> 2         2 -0.883        1      100
#> 3         3 -1.69         1      100
#> 4         4  0.000492     1      100
#> 5         5  1.17         1      100
#> 6         6 -0.738        1      100

# plot sample histogram
plot_sample_hist(pop, samples, height, c(10, 50, 100))
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

plot_sampling_hist(samples, height, c(10, 50, 100))
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Documentation

The official documentation is hosted on pkgdown. You can also refer to our html vignette.