R Programming Reference for AST 232

§ 1R and RStudio Setup

Install R from cran.r-project.org (the actual language) and RStudio Desktop from posit.co (a friendly editor around it). Open RStudio; the console at the bottom is where you type commands.

Working directory

R has a "current folder" it reads from and writes to. Check and change it with:

getwd()                              # where am I?
setwd("/Users/you/Desktop/ast232")   # go there (macOS / Linux)
setwd("C:/Users/you/Desktop/ast232")  # go there (Windows)

§ 2Basic Syntax and Assignment

x <- 5                # store 5 in x
y <- 3                # store 3 in y
x + y              # prints 8 to the console
z <- x + y          # store 8 in z, print nothing
print(z)           # prints 8 explicitly

§ 3Data Types

Factors are important in ANOVA

ANOVA treats a factor differently from a plain number. Always convert treatment / block / group columns to factors before aov():

treat <- c("V1", "V2", "V3", "V1", "V2", "V3")
treat <- factor(treat)
levels(treat)   # "V1" "V2" "V3"

§ 4Vectors

The vector is the fundamental R object. Even a single number is a vector of length 1.

Creating vectors

x <- c(1, 2, 3, 4, 5)              # combine
y <- 1:10                          # sequence 1 to 10
z <- seq(0, 1, by = 0.1)             # 0.0, 0.1, ..., 1.0
w <- seq(0, 100, length.out = 11)   # 0, 10, 20, ..., 100
r <- rep("A", times = 5)             # "A" "A" "A" "A" "A"
g <- rep(c("V1", "V2", "V3"), each = 4)   # V1 V1 V1 V1 V2 V2 ... V3 V3
h <- rep(1:4, times = 3)          # 1 2 3 4 1 2 3 4 1 2 3 4

Vector arithmetic is element wise

a <- c(1, 2, 3)
b <- c(10, 20, 30)
a + b       # 11 22 33
a * 2       # 2  4  6   (scalar broadcasts)
sqrt(a)     # 1.000 1.414 1.732

Handy summary functions

length(x)    # number of elements
sum(x)       # total
mean(x)      # mean
min(x); max(x)
range(x)     # c(min, max)
sort(x)      # sorted vector
rev(x)       # reverse
cumsum(x)    # cumulative sums, used in Life Table Tx
cumprod(x)   # cumulative products, used in Life Table lx

§ 5Data Frames

A data frame is a rectangular table where each column can be a different type. It is the R equivalent of a spreadsheet.

# Build one from vectors
age    <- c(18, 22, 30, 45)
gender <- c("F", "M", "F", "M")
height <- c(160, 175, 162, 180)

df <- data.frame(age, gender = factor(gender), height)
df
#   age gender height
# 1  18      F    160
# 2  22      M    175
# 3  30      F    162
# 4  45      M    180

Exploring a new data frame

head(df, n = 6)    # first 6 rows
tail(df, n = 6)    # last 6 rows
str(df)            # structure: types and first values
summary(df)        # quick summary per column
names(df)          # column names
nrow(df); ncol(df)
dim(df)            # c(nrow, ncol)

Adding a computed column

df$bmi <- 70 / (df$height / 100)^2    # assume weight 70 kg

§ 6Subsetting: [ ], [[ ]], $

Vector subsetting

x <- c(10, 20, 30, 40, 50)

x[3]              # 30 (third element)
x[2:4]            # 20 30 40
x[c(1, 3, 5)]      # 10 30 50 (integer index)
x[-1]             # drop the first element
x[x > 25]          # 30 40 50 (logical index)

Data frame subsetting

df[1, ]              # first row, all columns
df[, 2]              # second column, all rows (returns vector)
df[, "gender"]       # same, by name
df$gender           # same, dollar shortcut

df[df$age > 25, ]     # all rows where age > 25
df[df$gender == "F", c("age", "height")]   # female rows, two columns

[ ] vs [[ ]] in a nutshell

• x[1] returns a subset of the same type. On a data frame it returns a data frame.
• x[[1]] extracts the element itself. On a data frame it returns the column as a vector.
• x$name is a shortcut for x[["name"]]. Works on data frames and lists.

§ 7Reading and Writing Data

CSV files

# Read
dat <- read.csv("my_data.csv")
dat <- read.csv("my_data.csv", header = TRUE,
                stringsAsFactors = FALSE)

# Write
write.csv(dat, "output.csv", row.names = FALSE)

Excel files

install.packages("readxl")           # once
library(readxl)

dat <- read_excel("my_data.xlsx")
dat <- read_excel("my_data.xlsx", sheet = "Sheet2")

R's native format (.RData / .rds)

# .RData: save multiple objects together
save(dat, model, file = "analysis.RData")
load("analysis.RData")    # restores dat, model into workspace

# .rds: one object at a time, cleaner
saveRDS(model, "model.rds")
model <- readRDS("model.rds")

§ 8Control Flow: if, for, while

if / else

x <- 10
if (x > 5) {
  print("big")
} else if (x == 5) {
  print("exactly five")
} else {
  print("small")
}

Vectorised alternative: ifelse()

x <- c(2, 7, 4, 9)
ifelse(x > 5, "big", "small")
# "small" "big"   "small" "big"

for loop

for (i in 1:5) {
  print(i^2)
}

# Fill a vector inside a loop (pre allocate)
result <- numeric(10)
for (i in 1:10) {
  result[i] <- i^2
}

while loop

i <- 1
while (i <= 5) {
  print(i)
  i <- i + 1
}

§ 9Writing Your Own Functions

The syntax is function_name <- function(arg1, arg2, ...) { ... body ... }. The last expression evaluated is the return value (you can also use return()).

# A tiny function
square <- function(x) {
  x ^ 2
}
square(7)   # 49

# Default arguments
mean_or_median <- function(x, use_median = FALSE) {
  if (use_median) median(x) else mean(x)
}

# A function that returns multiple values via a named list
summary_stats <- function(x) {
  list(
    n       = length(x),
    mean    = mean(x),
    sd      = sd(x),
    range   = range(x)
  )
}
summary_stats(c(1, 4, 5, 6, 8))

§ 10The Apply Family

These functions apply another function across the elements of a vector, list, or margin of a matrix / data frame. They replace many for loops.

§ 11Common Statistical Functions

§ 12Distribution Functions

Every standard distribution comes with four functions, prefixed d (density), p (cumulative), q (quantile / inverse cumulative), and r (random draws).

How the four flavours differ

dnorm(1.96)              # density (height of the curve) at x = 1.96
pnorm(1.96)              # P(Z <= 1.96) = 0.975
qnorm(0.975)             # quantile: value with P below it of 0.975 = 1.96
rnorm(5, mean = 0, sd = 1) # 5 random draws

§ 13Formulas and aov()

A formula in R uses ~ to separate the response (left) from the predictors (right).

model <- aov(yield ~ treat + block, data = df)
summary(model)               # ANOVA table with F and p
residuals(model)             # residuals for diagnostics
fitted(model)                # fitted values
coef(model)                  # coefficients (contrast form)
TukeyHSD(model, "treat")     # pairwise comparison, family wise controlled

§ 14Plotting Basics

Base R plotting functions used in this course

Common styling arguments

plot(x, y,
     type = "o",           # "p" points, "l" lines, "b" both, "o" over-plotted
     pch  = 19,            # point shape (19 = filled circle)
     col  = "#2563eb",     # colour (name, hex, or rgb)
     lwd  = 2,             # line width
     xlab = "Age group",   # axis labels
     ylab = "Rate",
     main = "My plot",     # title
     xlim = c(0, 50),      # axis range
     ylim = c(0, 100))
points(x, y2, col = "red")     # overlay more points
lines(x, y3, col = "green")   # overlay a line
abline(h = 0, lty = 2)         # horizontal dashed reference line
legend("topright",
       legend = c("A", "B"),
       col    = c("red", "green"),
       lty    = 1, lwd = 2)

Multiple plots in one window

par(mfrow = c(1, 2))   # 1 row, 2 columns
qqnorm(res); qqline(res)
plot(fit, res); abline(h = 0, lty = 2)
par(mfrow = c(1, 1))   # reset back to single plot

§ 15Common Pitfalls

1. Character used where a factor is needed

If treat is a character vector (not a factor), aov() may still work but printing coefficients gets confusing. Convert with factor() before modelling.

2. Missing values silently propagate

mean(x) where x contains a single NA returns NA. Add na.rm = TRUE: mean(x, na.rm = TRUE).

3. Integer overflow

Very large integer arithmetic can overflow. Use doubles: multiply by 1.0 to force conversion.

4. Case sensitive names

Yield is not the same as yield. R errors will complain "object not found". Check your capitalisation.

5. Wrong assignment direction

5 -> x is legal R (right assignment) but almost always a mistake. Stick with <- or =.

6. Forgetting to reset par()

After par(mfrow = c(2, 2)), R stays in 2x2 layout until you set it back to c(1, 1) or restart R.

7. Working directory drift

If read.csv("data.csv") fails with "cannot open connection", your working directory is not what you think. Check with getwd(), fix with setwd() or (better) use RStudio's Session → Set Working Directory → To Source File Location.

8. Package not loaded

Installed once with install.packages("readxl"), but you still need library(readxl) at the top of every script that uses it.