new style for previous offerings

Author: Multipixels

data/gen-data.R

@@ -0,0 +1,15 @@
+
+# Generates 5 data files in the current folder
+
+for (i in 1:5) {
+  x <- rnorm(100)
+  y <- x + rnorm(100)
+  write.csv(
+    data.frame(
+      x = x,
+      y = y
+    ),
+    file = sprintf('data-file-%s.csv', i),
+    row.names = FALSE
+  )
+}

data/workshop-file.R

@@ -0,0 +1,286 @@
+
+### Atomic data types
+
+# Numbers
+x <- 1
+# arithmetic
+x + 1 # addition
+x - 1 # subtraction
+x * 2 # multiplication
+x / 2 # division
+x ** 2 # exponentiation
+1 / (2 - 0.5) # BEDMAS rules
+
+# Text (strings)
+greeting <- "hello"
+greeting <- 'hello'
+sentence <- 'He says "hello"'
+sentence <- "I'm busy now"
+
+# Booleans (true/false values)
+bool1 <- FALSE
+bool2 <- TRUE
+# Short forms (careful not to overwrite!)
+bool1 <- T
+bool2 <- F
+
+# Booleans as a result of comparisons
+x <- 2
+y <- 3
+x < y
+x <= y
+x > y
+x >= y
+x == y
+x != y
+
+# Boolean operators
+x <- TRUE
+y <- FALSE
+x & y # and
+x | y # or
+!y # not
+
+# Order of operations with Booleans
+
+# Suppose we want to know if we can publish our results. We can publish if the
+# results are significant or the study was preregistered, and if we have the
+# funds to publish open-access or the journal will waive open-access fee.
+# Suppose we have the funds to publish open-access even though the journal won't
+# waive the fee, but out results were not significant and the study was not
+# preregistered. Let's create some variables to represent the situation:
+funds <- TRUE # We have the funds
+waive <- FALSE # The journal won't waive the fee
+signif <- FALSE # Our results are not significant
+prereg <- FALSE # Our study was not preregistered
+# So, can we publish?
+can_publish <- funds | waive & signif | prereg
+print(can_publish)
+# Surprisingly, the above expression says we can! Why is this? It's because the
+# & operator takes precedence over the | operators, so it's evaluated first. The
+# expression then reduces to TRUE | FALSE | FALSE, which evaluates to TRUE
+# (regardless of which | operator is applied first). To ensure that the |
+# operators are applied first, we should use brackets:
+can_publish <- (funds | waive) & (signif | prereg)
+print(can_publish)
+# Now the result makes sense---we can't publish, but at least we know about
+# Boolean order of operations!
+
+### Collections
+
+# vectors
+
+x <- c(1, 2, 3, 4, 5)
+y <- c(6, 7, 8)
+z <- c(x, y)
+
+a <- c()
+b <- c(1, 2, 3)
+d <- c(a, b)
+
+a <- c(1, 2, 3, 'a', 'b', 'c') # Data types in vectors must all be the same
+print(a) # Note that numbers are converted to text
+
+# Storing values by name
+
+a <- c(first = 1, second = 2, third = 3, fourth = 4)
+print(a)
+
+# Accessing values in vectors
+
+a[3] # Single items
+a[2:4] # Ranges of items
+a[c(1, 3, 4)] # Specific items
+a[-2] # Excluding single items
+a[-2:-4] # Excluding ranges of items
+a[c(-1, -3, -4)] # Excluding specific items
+a['first'] # Accessing by name
+a[c('first', 'second')] # Accessing multiple by name
+
+# Storing mixed data types in lists
+
+a <- list(1, 2, 3, 'a', 'b', 'c')
+print(a)
+
+# Accessing actual items vs sub-lists
+
+a[[1]] # Item within list
+a[1] # Subsection of list
+
+# What's the difference?
+
+a[[1]] + 1
+a[1] + 1
+
+# Storing items by name
+a <- list(first = 1, second = 2, third = 3)
+print(a)
+a['first']
+a[['first']]
+a$first
+
+# Selecting multiple items
+a[[c('first', 'second')]] # Error!
+unlist(a[c('first', 'second')])
+
+### Conditionals
+
+# p values example:
+
+p <- 0.009
+
+if (p < 0.001) {
+  print('Our results are very significant!')
+} else if (p < 0.01) {
+  print('Our results are fairly significant')
+} else if (p < 0.05) {
+  print('Our results are technically significant')
+} else {
+  print('Our results are not significant')
+}
+
+if (p < 0.001) {
+  print('Our results are very significant!')
+}
+if (p < 0.01) {
+  print('Our results are fairly significant')
+}
+
+### Loops
+
+# Simulate example data
+n_data <- 100
+x <- rnorm(n_data)
+y <- x + rnorm(n_data)
+plot(y ~ x)
+
+# Compute a single bootstrap correlation
+n_sample <- 50
+sample_idx <- sample(1:n_data, n_sample, replace = TRUE)
+sampled_x <- x[sample_idx]
+sampled_y <- y[sample_idx]
+cor(sampled_x, sampled_y)
+
+# Do bootstrapped correlations in a loop
+n_iters <- 200
+iter_count <- 0
+cors <- c()
+repeat {
+  # Compute bootstrap correlation
+  sample_idx <- sample(1:n_data, n_sample, replace = TRUE)
+  sampled_x <- x[sample_idx]
+  sampled_y <- y[sample_idx]
+  curr_cor <- cor(sampled_x, sampled_y)
+  # Accumulate
+  cors <- c(cors, curr_cor)
+  # Keep count
+  iter_count <- iter_count + 1
+  if (iter_count == n_iters) {
+    break
+  }
+}
+
+# More succinctly using a while loop:
+iter_count <- 0
+cors <- c()
+while (iter_count < n_iters) {
+  sample_idx <- sample(1:n_data, n_sample, replace = TRUE)
+  sampled_x <- x[sample_idx]
+  sampled_y <- y[sample_idx]
+  curr_cor <- cor(sampled_x, sampled_y)
+  cors <- c(cors, curr_cor)
+  iter_count <- iter_count + 1
+}
+
+# Yet more succinctly using a for loop
+cors <- rep(NA, n_iters)
+for (iter_count in 1:n_iters) {
+  sample_idx <- sample(1:n_data, n_sample, replace = TRUE)
+  sampled_x <- x[sample_idx]
+  sampled_y <- y[sample_idx]
+  curr_cor <- cor(sampled_x, sampled_y)
+  cors[iter_count] <- curr_cor
+}
+
+# Example
+
+data.file.names <- c(
+  'data-file-1.csv',
+  'data-file-2.csv',
+  'data-file-3.csv',
+  'data-file-4.csv',
+  'data-file-5.csv'
+)
+
+cors <- c()
+for (data.file.name in data.file.names) {
+  data.file <- read.csv(data.file.name)
+  cors <- c(cors, xycor(data.file))
+}
+
+### Functions
+
+bootstrap_corr <- function(x, y, n_iters, n_samples) {
+  cors <- c()
+  for (i in 1:n_iters) {
+    sample_idx <- sample(1:n_data, n_sample, replace = TRUE)
+    cors[i] <- cor(x[sample_idx], y[sample_idx])
+  }
+  return(cors)
+}
+cors <- bootstrap_corr(x, y, 100, 50)
+hist(cors)
+
+# Default input values
+bootstrap_corr <- function(x, y, n_iters = 100, n_samples = 50) {
+  cors <- c()
+  for (i in 1:n_iters) {
+    sample_idx <- sample(1:n_data, n_sample, replace = TRUE)
+    cors[i] <- cor(x[sample_idx], y[sample_idx])
+  }
+  return(cors)
+}
+cors <- bootstrap_corr(x, y)
+hist(cors)
+
+# Functions within functions
+
+subsampled_cor <- function(x, y, n_samples) {
+  sample_idx <- sample(1:n_data, n_sample, replace = TRUE)
+  return(cor(x[sample_idx], y[sample_idx]))
+}
+
+bootstrap_corr <- function(x, y, n_iters = 100, n_samples = 50) {
+  cors <- c()
+  for (i in 1:n_iters) {
+    cors[i] <- subsampled_cor(x, y, n_samples) # !!
+  }
+  return(cors)
+}
+
+# Local vs global variables
+
+# Local variables do not exist outside functions
+create_txt <- function() {
+  txt <- 'abc'
+  print(txt)
+}
+create_txt()
+print(txt)
+
+# Local variables can have the same names as global variables
+txt2 <- 'hello'
+change_txt2 <- function() {
+  txt2 <- 'goodbye'
+  print(txt2)
+}
+print(txt2)
+change_txt2()
+print(txt2)
+
+# <<- can create global variables even within functions
+create_txt3 <- function() {
+  txt3 <<- 'abc'
+}
+create_txt3()
+print(txt3)

previousOffering.md

@@ -12,9 +12,12 @@ parent: Home
 # Previous Offerings
 Workshop live recordings are provided here. They are an alternative to the workshop content provided on the website.
 
-## October 2022
-<iframe height="416" width="100%" allowfullscreen frameborder=0 src="https://echo360.ca/media/60aeca91-1d04-493f-a94e-ee733efa3520/public"></iframe>
-[View original here.](https://echo360.ca/media/60aeca91-1d04-493f-a94e-ee733efa3520/public)
+<details markdown="1">
+<summary>December 2022</summary>
+<iframe height="416" width="100%" allowfullscreen frameborder=0 src="https://echo360.ca/media/ac3c8750-dd6f-46df-9db7-0a158476f41a/public"></iframe>
+[View original here.](https://echo360.ca/media/ac3c8750-dd6f-46df-9db7-0a158476f41a/public)
 
-<embed src="assets/docs/examplePDF.pdf" style="border:none;" width="100%" height="466px">
-[Download as PDF.](assets/docs/examplePDF.pdf)
+Access the workshop files here:
+- [gen-data.R](data/gen-data.R)
+- [workshop-file.R](data/workshop-file.R)
+</details>