Day 5 - R-eady to be Deployed!

By the end of today, you will be able to:

Data used: mpg, economics (bundled with ggplot2).

Setup

Start by creating a new qmd file, like we did yesterday. Write/paste all the code from today in the same file so that we can render at the end.

# Install if needed (run once):
# install.packages(c("tidyverse"))

library(tidyverse)   # loads ggplot2, dplyr, tidyr, readr, tibble, stringr, forcats, purrr
```{r}
# Quick peek at the data we'll use
head(mpg)
str(mpg)
```

Conditionals in Base R

Conditional statements let your code make decisions.

Comparison operators (what they do)

  • ==, !=: test equality / inequality (return TRUE/FALSE).
  • >, >=, <, <=: numeric comparisons.
  • &&, ||: element-wise logic on single TRUE/FALSE values (short-circuit). Use &/| for vectorized logic.
  • !: logical negation.
```{r}
x <- 10
y <- 20

x == y     # FALSE
x != y     # TRUE
x > 0      # TRUE
x >= 10    # TRUE
x < y      # TRUE
x <= y     # TRUE
```

if

  • Runs a code block only if the condition is TRUE.
```{r}
if (x > 0) {
  print("x is a positive number")
}
```

if ... else

  • Chooses one of two paths.
```{r}
if (x > y) {
  print("x is greater than y")
} else {
  print("y is greater than or equal to x")
}
```

Laddered else if

  • Chains multiple mutually exclusive tests.
```{r}
x <- y
if (x > y) {
  print("x is greater than y")
} else if (y > x) {
  print("y is greater than x")
} else {
  print("x and y are equal")
}
```

Nested if

  • Scope one decision inside another.
```{r}
x <- 20; y <- 10; z <- -30
if (x > y) {
  print("x is greater than y")
  if (x > z) {
    print("x is greater than y and z")
  }
}
```

Multiple conditions and the difference between &/| vs &&/||

```{r}
# For single TRUE/FALSE (first element only): &&, ||
if (x > y && x > z) {
  print("x is greater than both y and z")
}

# Vectorized comparisons across all elements: &, |
c(1,2,3) > 2 & c(10,20,30) > 15
```
Exercise

Write an if...else if...else ladder that prints "small", "medium", or "large" depending on the value of n relative to 10 and 100.

Descriptive Statistics & Aggregation (Base R)

Summarize the mpg dataset to understand central tendency, spread, and relationships.

Central tendency

  • mean(x): arithmetic average (sensitive to outliers).
  • median(x): middle value (robust to outliers).
```{r}
mean(mpg$displ)
median(mpg$displ)

# Note: base R's mode() returns the storage mode, not the statistical mode.
mode(mpg$displ)
```

Spread & quantiles

  • var(x), sd(x): variance and SD.
  • quantile(x, p): p-th quantile (e.g., 0.25 = Q1).
  • IQR(x): Q3 - Q1.
  • range(x), min(x), max(x).
```{r}
var(mpg$cty)
sd(mpg$cty)

quantile(mpg$cty, c(0.25, 0.75))
IQR(mpg$cty)

min(mpg$cty); max(mpg$cty)
range(mpg$cty)
```

Categorical summaries

  • table(x): counts by category.
  • unique(x): distinct values.
```{r}
table(mpg$drv)
unique(mpg$cyl)
```

Group-wise means (manual vs. aggregate())

  • aggregate(y ~ g, FUN, data): summary of y by groups g.
  • subset = lets you compute summaries on filtered rows.
```{r}
# Manual subsetting (repetitive)
mean(mpg$cty[mpg$cyl == 4])
mean(mpg$cty[mpg$cyl == 6])
mean(mpg$cty[mpg$cyl == 8])

# Using aggregate(): dv ~ iv
aggregate(cty ~ cyl, FUN = mean, data = mpg)

# Subset argument
aggregate(cty ~ cyl,
          FUN = mean,
          subset = cty < 15,
          data = mpg)
```

Quick summaries (summary(), by())

  • summary(df) gives per-column summaries.
  • by(df, group, summary) applies summary() within each group.
```{r}
summary(mpg)
summary(mpg$displ)

by(mpg, mpg$cyl, summary)
```

Contingency tables & proportions (table(), xtabs(), prop.table())

  • xtabs(~ a + b): formula interface for contingency tables.
  • prop.table(tab, margin): convert counts to proportions (by rows 1 or columns 2).
```{r}
# Counts
table(mpg$cyl, mpg$year)
xtabs(~ mpg$cyl + mpg$year)

# Relative frequencies
prop.table(table(mpg$year, mpg$class))
prop.table(table(mpg$year, mpg$class), 1)          # by row
round(prop.table(table(mpg$year, mpg$class), 2), 2) # by column (rounded)
round(prop.table(table(mpg$year, mpg$class), 2), 2) * 100
```

Mosaic plot

```{r}
mosaicplot(table(mpg$year, mpg$class),
           color = TRUE,
           xlab = "Year",
           ylab = "Class")
```
Exercise

Compute the median highway mileage (hwy) by number of cylinders (cyl) with aggregate(). Compare it to the corresponding boxplots in Section 4.6.

Data Wrangling with dplyr (Pipes, select, filter, mutate, summarise, joins)

The dplyr verbs make data manipulation fast and expressive. They work best with the pipe |> (base R) or %>% (magrittr). We’ll use |> here.

dplyr is a part of tidyverse universe. Tidyverse is an opiniated approach to data handling. What this means is that the packages and practices in this approach presuppose a particular format of data. The data either needs to be in the particular format or be transformed to one, for this approach to work efficiently. In practice, it is quite intuitive and simple.

There are three interrelated rules of tidy data are:

1- Each variable must have its own column.
2- Each observation must have its own row.
3- Each value must have its own cell.

You can read more about tidy data here.

Pipes

# Without pipe
head(arrange(select(mpg, manufacturer, model, hwy), desc(hwy)), 5)

# With pipe (clearer)
mpg |>
  select(manufacturer, model, hwy) |>
  arrange(desc(hwy)) |>
  head(5)
# A tibble: 5 × 3
  manufacturer model        hwy
  <chr>        <chr>      <int>
1 volkswagen   jetta         44
2 volkswagen   new beetle    44
3 volkswagen   new beetle    41
4 toyota       corolla       37
5 honda        civic         36
# A tibble: 5 × 3
  manufacturer model        hwy
  <chr>        <chr>      <int>
1 volkswagen   jetta         44
2 volkswagen   new beetle    44
3 volkswagen   new beetle    41
4 toyota       corolla       37
5 honda        civic         36

select() (choose columns)

mpg |>
  select(manufacturer, model, hwy)

mpg |>
  select(starts_with("d")) |>
  head(3)

# Rename while selecting
mpg |>
  select(make = manufacturer, model, hwy) |>
  head(3)

# Move columns around
mpg |>
  relocate(hwy, .before = displ) |>
  head(3)
# A tibble: 234 × 3
   manufacturer model        hwy
   <chr>        <chr>      <int>
 1 audi         a4            29
 2 audi         a4            29
 3 audi         a4            31
 4 audi         a4            30
 5 audi         a4            26
 6 audi         a4            26
 7 audi         a4            27
 8 audi         a4 quattro    26
 9 audi         a4 quattro    25
10 audi         a4 quattro    28
# ℹ 224 more rows
# A tibble: 3 × 2
  displ drv  
  <dbl> <chr>
1   1.8 f    
2   1.8 f    
3   2   f    
# A tibble: 3 × 3
  make  model   hwy
  <chr> <chr> <int>
1 audi  a4       29
2 audi  a4       29
3 audi  a4       31
# A tibble: 3 × 11
  manufacturer model   hwy displ  year   cyl trans      drv     cty fl    class 
  <chr>        <chr> <int> <dbl> <int> <int> <chr>      <chr> <int> <chr> <chr> 
1 audi         a4       29   1.8  1999     4 auto(l5)   f        18 p     compa…
2 audi         a4       29   1.8  1999     4 manual(m5) f        21 p     compa…
3 audi         a4       31   2    2008     4 manual(m6) f        20 p     compa…

filter() (keep rows)

mpg |>
  filter(class == "suv", hwy >= 20) |>
  head(5)

mpg |>
  filter(manufacturer %in% c("toyota", "honda")) |>
  count(manufacturer)
# A tibble: 5 × 11
  manufacturer model       displ  year   cyl trans drv     cty   hwy fl    class
  <chr>        <chr>       <dbl> <int> <int> <chr> <chr> <int> <int> <chr> <chr>
1 chevrolet    c1500 subu…   5.3  2008     8 auto… r        14    20 r     suv  
2 chevrolet    c1500 subu…   5.3  2008     8 auto… r        14    20 r     suv  
3 jeep         grand cher…   3    2008     6 auto… 4        17    22 d     suv  
4 jeep         grand cher…   4    1999     6 auto… 4        15    20 r     suv  
5 nissan       pathfinder…   4    2008     6 auto… 4        14    20 p     suv  
# A tibble: 2 × 2
  manufacturer     n
  <chr>        <int>
1 honda            9
2 toyota          34

mutate() (create/modify columns)

mpg |>
  mutate(
    hwy_kmpl = hwy * 0.425144,           # convert mpg to km/L (approx)
    efficiency = case_when(
      hwy >= 30 ~ "high",
      hwy >= 20 ~ "medium",
      TRUE ~ "low"
    )
  ) |>
  select(manufacturer, model, hwy, hwy_kmpl, efficiency) |>
  head(6)
# A tibble: 6 × 5
  manufacturer model   hwy hwy_kmpl efficiency
  <chr>        <chr> <int>    <dbl> <chr>     
1 audi         a4       29     12.3 medium    
2 audi         a4       29     12.3 medium    
3 audi         a4       31     13.2 high      
4 audi         a4       30     12.8 high      
5 audi         a4       26     11.1 medium    
6 audi         a4       26     11.1 medium

arrange() (sort rows)

mpg |>
  arrange(manufacturer, desc(hwy)) |>
  head(6)
# A tibble: 6 × 11
  manufacturer model      displ  year   cyl trans  drv     cty   hwy fl    class
  <chr>        <chr>      <dbl> <int> <int> <chr>  <chr> <int> <int> <chr> <chr>
1 audi         a4           2    2008     4 manua… f        20    31 p     comp…
2 audi         a4           2    2008     4 auto(… f        21    30 p     comp…
3 audi         a4           1.8  1999     4 auto(… f        18    29 p     comp…
4 audi         a4           1.8  1999     4 manua… f        21    29 p     comp…
5 audi         a4 quattro   2    2008     4 manua… 4        20    28 p     comp…
6 audi         a4           3.1  2008     6 auto(… f        18    27 p     comp…

group_by() + summarise() (grouped summaries)

mpg |>
  group_by(cyl) |>
  summarise(
    mean_cty = mean(cty),
    median_hwy = median(hwy),
    n = dplyr::n()
  )

mpg |>
  group_by(drv) |>
  summarise(across(c(cty, hwy), list(mean = mean, sd = sd)))
# A tibble: 4 × 4
    cyl mean_cty median_hwy     n
  <int>    <dbl>      <dbl> <int>
1     4     21.0         29    81
2     5     20.5         29     4
3     6     16.2         24    79
4     8     12.6         17    70
# A tibble: 3 × 5
  drv   cty_mean cty_sd hwy_mean hwy_sd
  <chr>    <dbl>  <dbl>    <dbl>  <dbl>
1 4         14.3   2.87     19.2   4.08
2 f         20.0   3.63     28.2   4.21
3 r         14.1   2.22     21     3.66

distinct(), slice(), sampling

# Unique combinations of columns
mpg |>
  distinct(manufacturer, class) |>
  head(10)

# Take specific rows by position
mpg |>
  slice(1:5)

# Random samples
set.seed(123)
mpg |>
  slice_sample(n = 5)
# A tibble: 10 × 2
   manufacturer class  
   <chr>        <chr>  
 1 audi         compact
 2 audi         midsize
 3 chevrolet    suv    
 4 chevrolet    2seater
 5 chevrolet    midsize
 6 dodge        minivan
 7 dodge        pickup 
 8 dodge        suv    
 9 ford         suv    
10 ford         pickup 
# A tibble: 5 × 11
  manufacturer model displ  year   cyl trans      drv     cty   hwy fl    class 
  <chr>        <chr> <dbl> <int> <int> <chr>      <chr> <int> <int> <chr> <chr> 
1 audi         a4      1.8  1999     4 auto(l5)   f        18    29 p     compa…
2 audi         a4      1.8  1999     4 manual(m5) f        21    29 p     compa…
3 audi         a4      2    2008     4 manual(m6) f        20    31 p     compa…
4 audi         a4      2    2008     4 auto(av)   f        21    30 p     compa…
5 audi         a4      2.8  1999     6 auto(l5)   f        16    26 p     compa…
# A tibble: 5 × 11
  manufacturer model       displ  year   cyl trans drv     cty   hwy fl    class
  <chr>        <chr>       <dbl> <int> <int> <chr> <chr> <int> <int> <chr> <chr>
1 pontiac      grand prix    5.3  2008     8 auto… f        16    25 p     mids…
2 toyota       toyota tac…   4    2008     6 auto… 4        16    20 r     pick…
3 toyota       4runner 4wd   4.7  2008     8 auto… 4        14    17 r     suv  
4 audi         a4 quattro    3.1  2008     6 auto… 4        17    25 p     comp…
5 toyota       corolla       1.8  1999     4 auto… f        24    33 r     comp…

Joins (combine tables by keys)

a <- tibble(id = c(1,2,3), val = c("A","B","C"))
b <- tibble(id = c(2,3,4), note = c("two","three","four"))

left_join(a, b, by = "id")
inner_join(a, b, by = "id")
full_join(a, b, by = "id")
anti_join(a, b, by = "id")  # rows in a without matches in b
semi_join(a, b, by = "id")  # rows in a that have a match in b
# A tibble: 3 × 3
     id val   note 
  <dbl> <chr> <chr>
1     1 A     <NA> 
2     2 B     two  
3     3 C     three
# A tibble: 2 × 3
     id val   note 
  <dbl> <chr> <chr>
1     2 B     two  
2     3 C     three
# A tibble: 4 × 3
     id val   note 
  <dbl> <chr> <chr>
1     1 A     <NA> 
2     2 B     two  
3     3 C     three
4     4 <NA>  four 
# A tibble: 1 × 2
     id val  
  <dbl> <chr>
1     1 A    
# A tibble: 2 × 2
     id val  
  <dbl> <chr>
1     2 B    
2     3 C

Reshaping with tidyr (bonus)

# Wide -> Long
wide <- tibble(id = 1:3, y_2020 = c(1,2,3), y_2021 = c(4,5,6))
long <- wide |>
  pivot_longer(starts_with("y_"), names_to = "year", values_to = "y")
long

# Long -> Wide
long |>
  pivot_wider(names_from = year, values_from = y)
# A tibble: 6 × 3
     id year       y
  <int> <chr>  <dbl>
1     1 y_2020     1
2     1 y_2021     4
3     2 y_2020     2
4     2 y_2021     5
5     3 y_2020     3
6     3 y_2021     6
# A tibble: 3 × 3
     id y_2020 y_2021
  <int>  <dbl>  <dbl>
1     1      1      4
2     2      2      5
3     3      3      6
Quick Cheatsheet
  • Columns: select(), rename(), relocate()
  • Rows: filter(), arrange(), slice()
  • New columns: mutate(), if_else(), case_when()
  • Groups: group_by(), summarise(), across()
  • Uniques & counts: distinct(), count()
  • Joins: left_join(), inner_join(), full_join(), anti_join(), semi_join()
  • Reshape: pivot_longer(), pivot_wider()

Grammar of Graphics with ggplot2

ggplot is the R functional form of the Grammar of Graphics

But what does it mean?

The Grammar of Graphics was introduced to describe the deep structure underlying statistical graphics. Wilkinson (2005) formalized the idea of what a “statistical graphic” is; Wickham (2010) adapted that grammar into a layered system in R. In short, a statistical plot maps data to aesthetic attributes (e.g., colour, shape, size) of geometric objects (e.g., points, lines, bars), may include statistical transformations, is drawn on a coordinate system, and can be faceted to repeat the same plot over subsets of the data. It’s the combination of these independent components that makes a graphic.

ggplot2 is built on the idea: any plot can be expressed from the same set of components—a dataset, a coordinate system, and geoms (visual marks). The key to understanding ggplot2 is thinking in layers, much like stacked layers in image editors such as Photoshop/Illustrator/Inkscape.

The guiding principle in ggplot2 is to build from foundational components (data, global mappings) to specific components (geoms, stats, scales, guides), connecting layers with +. Think of each layer as placed on the same canvas; together they compose the plot. The general structure is:

ggplot(data = <DATA>, mapping = aes(<X>, <Y>, <global aesthetics>)) +
  geom_<TYPE>(aes(<layer-specific aesthetics>), <layer options>) +
  stat_<TRANSFORM>(...) +
  scale_<...>(...) +
  coord_<...>(...) +
  facet_<...>(...) +
  labs(title = "...", subtitle = "...", x = "...", y = "...") +
  theme(...)

The ggplot2 package builds plots by layering components: data, aesthetic mappings, geoms, statistics, scales, coordinates, and facets.

You can read more about ggplot here.

First plot

ggplot(mpg, aes(x = displ, y = hwy)) +
  geom_point()

Compare to base R:

plot(x = mpg$displ, y = mpg$hwy)

Aesthetics: colour, size, shape, alpha

ggplot(mpg, aes(displ, hwy, colour = manufacturer)) + geom_point()
ggplot(mpg, aes(displ, hwy, colour = manufacturer, size = cyl)) + geom_point()
ggplot(mpg, aes(displ, hwy, shape = factor(year))) + geom_point()
ggplot(mpg, aes(displ, hwy, alpha = cyl)) + geom_point()

Fixed aesthetics (outside aes()):

ggplot(mpg, aes(displ, hwy)) + geom_point(aes(colour = "blue"))
ggplot(mpg, aes(displ, hwy)) + geom_point(colour = "blue")

Faceting

ggplot(mpg, aes(displ, hwy)) +
  geom_point() +
  facet_wrap(~ class)

Distributions: histogram, frequency polygon, density, combo

ggplot(mpg, aes(hwy)) + geom_histogram()
ggplot(mpg, aes(hwy)) + geom_freqpoly()

# Change binwidth
ggplot(mpg, aes(hwy)) + geom_histogram(binwidth = 2)
ggplot(mpg, aes(hwy)) + geom_freqpoly(binwidth = 2)

# Density
ggplot(mpg, aes(hwy)) + geom_density()

# Overlay histogram + density and add a mean line
p <- ggplot(mpg, aes(hwy)) +
  geom_histogram(aes(y = ..density..), colour = "black", fill = "white") +
  geom_density(alpha = .2, fill = "#FF6666")

p + geom_vline(aes(xintercept = mean(hwy)), linetype = "dashed", size = 1)
`stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

`stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
ℹ Please use `linewidth` instead.
Warning: The dot-dot notation (`..density..`) was deprecated in ggplot2 3.4.0.
ℹ Please use `after_stat(density)` instead.
`stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Box, violin, and jitter

ggplot(mpg, aes(drv, hwy)) + geom_boxplot()
ggplot(mpg, aes(drv, hwy)) + geom_jitter()
ggplot(mpg, aes(drv, hwy)) + geom_violin()

Bars: counts vs. pre-summarised values

# Counts from raw rows
ggplot(mpg, aes(manufacturer)) + geom_bar()

# Pre-summarised data
drugs <- data.frame(
  drug = c("a", "b", "c"),
  effect = c(4.2, 9.7, 6.1)
)
ggplot(drugs, aes(drug, effect)) + geom_bar(stat = "identity")

Lines & paths (time series)

ggplot(economics, aes(date, unemploy / pop)) + geom_line()
ggplot(economics, aes(date, uempmed)) + geom_line()

str(economics)

spc_tbl_ [574 × 6] (S3: spec_tbl_df/tbl_df/tbl/data.frame)
 $ date    : Date[1:574], format: "1967-07-01" "1967-08-01" ...
 $ pce     : num [1:574] 507 510 516 512 517 ...
 $ pop     : num [1:574] 198712 198911 199113 199311 199498 ...
 $ psavert : num [1:574] 12.6 12.6 11.9 12.9 12.8 11.8 11.7 12.3 11.7 12.3 ...
 $ uempmed : num [1:574] 4.5 4.7 4.6 4.9 4.7 4.8 5.1 4.5 4.1 4.6 ...
 $ unemploy: num [1:574] 2944 2945 2958 3143 3066 ...

Axes and limits

ggplot(mpg, aes(cty, hwy)) + geom_point(alpha = 1/3)

ggplot(mpg, aes(cty, hwy)) +
  geom_point(alpha = 1/3) +
  xlab("city driving (mpg)") +
  ylab("highway driving (mpg)")

ggplot(mpg, aes(cty, hwy)) + geom_point(alpha = 1/3) + xlab(NULL) + ylab(NULL)

ggplot(mpg, aes(drv, hwy)) + geom_jitter(width = 0.25)

ggplot(mpg, aes(drv, hwy)) +
  geom_jitter(width = 0.25) +
  xlim("f", "r") +
  ylim(20, 30)

ggplot(mpg, aes(drv, hwy)) +
  geom_jitter(width = 0.25, na.rm = TRUE) +
  ylim(NA, 30)

Warning: Removed 139 rows containing missing values or values outside the scale range
(`geom_point()`).

Annotations & math expressions

ggplot(mpg, aes(displ, hwy)) +
  geom_point(aes(colour = factor(cyl))) +
  labs(
    x = "Engine displacement (litres)",
    y = "Highway miles per gallon",
    colour = "Number of cylinders",
    title = "Mileage by engine size and cylinders",
    subtitle = "Source: http://fueleconomy.gov"
  )

values <- seq(from = -2, to = 2, by = .01)
df <- data.frame(x = values, y = values^3)
ggplot(df, aes(x, y)) +
  geom_path() +
  labs(y = quote(f(x) == x^3))

Saving plots

# Save a plot object to disk
p <- ggplot(df, aes(x, y)) +
  geom_path() +
  labs(y = quote(f(x) == x^3))
ggsave("plot.png", p, width = 5, height = 5)

# Or specify a full path you control:
# ggsave("path/to/your/folder/plot.png", p, width = 5, height = 5)

Wrap-up

  • Base R conditionals are the building blocks of program logic.
  • Base R provides quick descriptive and grouped summaries (aggregate(), by(), table(), xtabs(), prop.table()).
  • dplyr enables readable, chained transformations: think rows (filter()), columns (select(), mutate()), groups (group_by() + summarise()), and joins.
  • ggplot2 enables layered, expressive visualizations; think in layers and mappings.

Practice: Using mpg, (1) create efficiency with mutate() and case_when(), (2) compute average hwy by class with group_by()/summarise(), and (3) plot hwy vs displ coloured by class with a linear smoother.