In trinker/reports: An R Package to Assist in the Workflow of Writing Academic Articles and Other Reports

.segue .dark .nobackground

Introduction

opts_chunk$set(prompt = FALSE, tidy = FALSE, message = F, comment = NA)
library(lattice); library(latticeExtra); library(mosaic)
trellis.par.set(theme = theEconomist.theme(with.bg = TRUE, box = 'transparent'))
lattice.options(theEconomist.opts())
# knit_hooks$set(output = function(x, options){
#   gsub('\n\n', '\n', x)
# })
knit_hooks$set(document = function(x){
  gsub("```\n*```r*\n*", "", x)
})

--- .segue .dark .nobackground

Slides

*** =pnotes

The classical environment for teaching has been the ubiqutous slide deck, without which no class, online or offline is considered complete. It serves dual roles, as a synchronoous teaching aid in class that helps anchor a lecture, and as an asynchronous record of what was taught, allowing students to revisit the contents later.

---

Introduction to ggplot2

# install ggplot2 if you have not already done so
# install.packages('ggplot2')
require(ggplot2)
qplot(wt, mpg, data = mtcars)

*** =pnotes

In this tutorial, you will learn the basics of ggplot2, an R package that makes it easy to create stunning statistical graphics. The first step is to install ggplot2

*** =spnotes

For example, here is how an introductory slide on a ggplot2 tutorial might look like. There are two shortcomings of a slides-only environment.

1. It would work great in class, with an instructor lecturing along. But as a standalone, it falls short. This concern can be partly alleviated by sharing speaker notes. You can press p to see the speaker notes for this slide. However, it is still not efficient.

2. To follow along, a student would need to copy the code from the slide, paste it into his/her IDE, run the code to see the results and then return back to the slide deck. I believe a lot of back and forth between slide and IDE is highly distracting to the learning process.

---

Data Frames

We will work with the built-in dataset mtcars

head(mtcars, 3)

You can inspect the data set in many ways.

head(mtcars)       # view first six rows
head(mtcars, 10)   # view first ten rows
tail(mtcars, 20)   # view last twenty rows
NROW(mtcars)       # view number of rows 
NCOL(mtcars)       # view number of columns
dim(mtcars)        # view rows x columns

---

Subsetting Data

Very often you might want to work with a smaller subset of the data. R makes it very easy to select subsets by combining conditional operations. The main operations are

== : equal to
!= : not equal to
>= : greater or equal to
<= : lesser or equal to
&  : and
|  : or

Here are some examples of how to choose specific subsets of the mtcars data set.

subset(mtcars, gear == 4)           # four gears
subset(mtcars, cyl != 2)            # not 2 cylinders
subset(mtcars, mpg > 20)            # mpg more than 20
subset(mtcars, mpg > 20 & wt < 20)  # mpg > 20 AND wt < 20
subset(mtcars, mpg > 20 | wt < 20)  # mpg > 20 OR wt < 20

--- .segue bg:lightblue .nobackground

Data Structures

--- &opencpu2 sno:2

Vector

A vector is a collection of items (or elements) that are all of the same type (integers, numbers, strings, true/false values). You can create a vector by simply passing a comma separated set of items to the c() function, which concatenates them into a vector.

num_vec  = c(1, 2, 3, 4, 5)      
char_vec = c('Ross', 'Chandler') 
logi_vec = c(TRUE, FALSE, FALSE) 

*** =question

Create a vector consisting of abbreviated weekdays (Mon, Tue etc.) and assign it to the variable weekdays

--- &opencpu2 sno:3

Vector

A vector has two key properties: length and mode.

num_vec <- c(1, 2, 3, 4, 5)
length(num_vec)
mode(num_vec)

*** =question

Task: Create a vector x <- c('alpha', 'beta', 2). Can you guess what its mode and length would be? Use the functions learnt above to figure it out for yourself!

*** =hint

A little surprised that mode(x) returns character. Well, recall that a vector by definition consists of elements of the same type. So, R does what is called type coercion and coerces the number 2 to a character string "2".

--- &opencpu2 sno:4

List

A list, like a vector is a collection of items. But unlike a vector, the items can be of different types. Here is an example of a list.

list(c(1, 2, 3), TRUE, 'gamma')

Note that it consists of three items, the first is a vector (of characters), the second is a logical value, while the third is a numeric.

*** =question

Verify for yourself the mode of each item of the list. You will need to assign the list to a variable first.

--- &opencpu sno:5

Exercise 1

Create a vector representing the radii of three circles with lengths 5, 10, and 20. Use * and the built-in constant pi to compute the areas of the three circles. Then subtract 2.1 from each radius and recompute the areas.

*** =hint

myradii  <- c(5, 10, 20)
pi * myradii^2
pi * (myradii - 2.1)^2

--- .segue .dark .nobackground

Control Structures

---

If-Then-Else

The if-then-else

day = "Sunday"
if (day == 'Sunday'){
  print("Sunday")
} else {
 print("Not a Sunday")
}

--- .segue .dark .nobackground

Loops

---

For Loop

A for loop can be used to carry out a computational task repeatedly.

x <- c(1, 2, 3, 4, 5)
mysum = 0
for (i in 1:5){
  mysum = mysum + x[i]
}
mysum

--- bg:lightgreen

While Loop

--- bg:lightblue

Repeat Loop

---

--- .quiz &radio

Question

How many different ways can we pick two marbles from a bag containing three marbles, colored red, blue and green?

1. 3
2. 6
3. 9

*** .explanation

This is the answer. Does math work?

$$\alpha = \beta$$

--- .quiz &text ans:10

Text Question

This question requires a text input as answer. What is 5 + 5?

*** .hint

Think addition!

*** .explanation

It is trivial

---

What is Probability?

Probability is the

--- .segue .dark .nobackground

Birthday Problem

--- .fill .nobackground bg:url(http://goo.gl/MfFa)

---

Hypothesise

Let us start by hypothesizing what the answer could be.

--- .nobackground

Simulate

We will use simulation to answer our question. Click here and open in a new tab.

---

Use R

There are two built in functions pbirthday and qbirthday which calculate the probability of a coincidence and the minimum number of people required for a given probability of coincidence. You can get more details by typing ? pbirthday in RStudio

Here are some hints on how to apply these functions

``` {r eval = F} pbirthday(n = 60) # prob. at least two people share a bday pbirthday(n = 60, coincident = 3) # prob. at least three people share a bday qbirthday(prob = 0.5, class = 365) # min. number required for a 50% prob. of common bday qbirthday(prob = 0.8, class = 30) # min. number required for a 80% prob. of common month

--- .quiz &radio

## Question 1

What is the minimum number of people required for a 99% probability of a common birthday? Choose the closest possible answer.

1. 30
2. 40
3. _60_
4. > 100

*** =pnotes

```r
qbirthday(0.99, classes = 365)

--- .quiz &text ans:57

Question 2

What is the minimum number of people required for a 99% probability of a common birthday?

--- .segue .dark .nobackground

Let's Make a Deal

---

Motivate

Should we base decisions on our intuition? Let's check your intuition to determine a probability with Monty Hall's Let's Make A Deal.

Monty Hall's game show Let's Make A Deal was a popular seventies game show. Nearly the entire audience dresses up in costumes hoping that Monty Hall would select them out of the crowd and offer them a chance to win a fabulous prize. He might offer $100 for every paper clip in your possession!

Here is the situation: Suppose you are given three doors to choose from. Behind one door there is a big prize (a car) and behind the other two, there are goats. Only Monty Hall knows which door has the prize. You are asked to select a door, and then Monty opens a different door, showing you a goat behind it. Then you are asked the big question: Do you want to stay with your original door, or switch to a different door?

Question: Do you have a higher chance of winning the prize if you stay with your first door selection, or switch to the remaining door?

---

Hypothesize

Let us start by hypothesizing what the answer could be.

--- .fill

Simulate

---

Ponder

--- .segue bg:darkslategrey

One Son Policy

---

Motivate

In the early 1990’s China considered adopting a "one son" policy, to help reduce their birthrate by allowing families to keep having children until they had a son. Under this plan a family has a child. If it is a son, they stop having children. If it is a daughter, they can try again. They can keep trying until they have a son and then they stop having children.

In small groups, discuss the following questions:

1. If a country adopted this “one son” policy, what would you expect the average number of children to be for a family, and why?
2. What would you expect the ratio of boys to girls to be (more girls than boys, more boys than girls, equal numbers of girls and boys)? Why?

---

Think

Working in teams, take a yogurt container that contains two slips of paper. One is labeled B for boy. One is labeled G for girl. You are going to randomly draw one slip of paper from the container and that will be the first child born in a simulated family. If you draw a B, then stop. Enter the data on the chart below. If it is a G, draw again. Keep drawing until you draw a B, then stop and enter the data on the chart below. Repeat this process for 5 simulated families. For example: GB, B, GGB, B, GB, etc.

---

Hypothesize

---

Simulate

--- .segue .dark .nobackground

Flip Coins with R

---

Flipping a Coin

We can use R to simulate the physical process of flipping a coin. You will need to install the mosaic package, which can be done by typing install.packages('mosaic') in RStudio

Let us start by flipping a coin 5 times

``` {r flip-1-5, comment = NA} rflip(5)

---

## Flipping Multiple Times

We can flip a coin 5 times and automatically count the number of heads tossed

``` {r flip-1-5-count}
nflip(5)

---

Flipping a Coin 500 times

We can now repeat the process of flipping 5 coins, 500 times and plot a histogram of the distribution.

{r flip-5-500, fig.width = 5, fig.height = 5, fig.align = 'center'} do500 = do(500) * rflip(5) # flip 5 coins, 500 times histogram(~ heads, data = do500, nint = 6) # plot a histogram of number of heads

*** =pnotes You can type View(do500) in RStudio if you want to take a look at the raw data generated.

trinker/reports documentation built on May 31, 2019, 9:51 p.m.