08 · Basic Plotting¶

`plot()` — R's built-in plotting function¶

Base R ships with a capable plotting system, no packages required. plot() adapts its behavior to the type of data you give it.

ages <- c(23, 25, 22, 34, 28, 41, 19, 30)
scores <- c(88, 92, 79, 65, 85, 70, 95, 81)

plot(ages, scores)
# Opens a plot window (or saves to a device) with a scatter plot:
# ages on the x-axis, scores on the y-axis, one point per pair

Customizing a scatter plot¶

plot(ages, scores,
     main = "Age vs. Test Score",
     xlab = "Age",
     ylab = "Score",
     col = "steelblue",
     pch = 16)          # pch = point character; 16 is a solid filled circle

Argument	Meaning
`main`	Plot title
`xlab` / `ylab`	Axis labels
`col`	Point/line color
`pch`	Point shape (0-25 are built-in symbols)
`type`	`"p"` points (default), `"l"` line, `"b"` both
`xlim` / `ylim`	Axis ranges, e.g. `xlim = c(0, 50)`

Line plots¶

days <- 1:7
temps <- c(15, 17, 16, 19, 22, 21, 18)

plot(days, temps, type = "l", col = "darkred", lwd = 2,
     main = "Temperature Over a Week", xlab = "Day", ylab = "°C")

type = "l" connects points with a line instead of drawing separate markers; lwd controls line width.

Bar charts — `barplot()`¶

sales <- c(120, 90, 150, 80)
names(sales) <- c("Q1", "Q2", "Q3", "Q4")

barplot(sales,
        main = "Quarterly Sales",
        ylab = "Units Sold",
        col = "seagreen")

barplot() expects a vector (ideally named, as above, so the names become the bar labels automatically).

Histograms — `hist()`¶

set.seed(42)                       # makes the "random" data reproducible
exam_scores <- rnorm(200, mean = 75, sd = 10)   # 200 values from a normal distribution

hist(exam_scores,
     main = "Distribution of Exam Scores",
     xlab = "Score",
     col = "cornflowerblue",
     breaks = 15)                  # roughly how many bars to draw

A histogram groups continuous values into bins and shows how many fall into each — the standard first look at any single numeric column's distribution.

Boxplots — `boxplot()`¶

group_a <- c(88, 92, 79, 65, 85)
group_b <- c(70, 95, 60, 55, 80)

boxplot(group_a, group_b,
        names = c("Group A", "Group B"),
        main = "Score Comparison",
        ylab = "Score",
        col = c("lightblue", "lightpink"))

A boxplot summarizes a distribution's median, quartiles, and outliers in one compact shape — useful for comparing two or more groups at a glance.

Saving a plot to a file¶

png("scatter.png", width = 800, height = 600)   # open a PNG file device
plot(ages, scores, main = "Age vs. Score")
dev.off()                                        # close the device -- writes the file

Every plotting call between opening a device (png(), pdf(), jpeg()) and dev.off() gets drawn into that file instead of the interactive plot window. Forgetting dev.off() is a common mistake — the file stays empty/locked until you close the device.

Multiple plots in one figure¶

par(mfrow = c(1, 2))     # 1 row, 2 columns of plots
plot(ages, scores, main = "Scatter")
hist(exam_scores, main = "Histogram")
par(mfrow = c(1, 1))     # reset back to a single plot per figure

Base plotting cheat sheet¶

Function	Use for
`plot(x, y)`	Scatter plot of two numeric vectors
`plot(x, y, type = "l")`	Line plot
`barplot(x)`	Bar chart of a (named) vector
`hist(x)`	Distribution of one numeric vector
`boxplot(x, y, ...)`	Compare distributions across groups
`png()` / `dev.off()`	Save a plot to a file
`par(mfrow = c(r, c))`	Arrange multiple plots in a grid

Base R plotting is quick and dependency-free, which is why it's introduced first. Level 2 introduces ggplot2, the more expressive and widely used plotting package for anything beyond a quick look at your data.

How It Actually Works¶

Base R plotting (plot(), hist(), abline(), ...) is imperative and stateful: each call draws directly onto the currently active graphics device (a bitmap or vector canvas R maintains internally) and mutates it in place. plot() doesn't just draw points — it resets the device, computes axis ranges from your data, and pushes a new "plotting region" onto R's internal graphics state stack. Every subsequent call like abline() or points() you make afterward doesn't redraw the whole chart; it layers more ink onto the same device using the axis coordinate system that plot() already established — which is exactly why call order matters and why you can't easily "undo" one layer without restarting from plot().

Under the hood, each high-level plotting function ultimately calls into R's grid/graphics engine (a C-level device driver interface), which translates your plotting calls into device-specific drawing operations — pixel writes for PNG, vector path commands for PDF/SVG. This is also why resizing an RStudio plot pane after the fact can shift text and point sizes: the device is re-rendered at the new dimensions from the same recorded calls, not simply scaled as an image.

🔀 See this in another language¶

Exercise¶

Using the exam_scores vector generated above (rnorm(200, mean = 75, sd = 10) with set.seed(42)), produce three plots: a histogram with 20 breaks and a title, a boxplot with a y-axis label, and a line plot of sort(exam_scores) (the sorted scores) to visualize the distribution's shape. Save the histogram to a file called scores_hist.png using png()/dev.off().