Hall of Fame

The evolution of a beeswarm

A categorical dot plot with a mean marker and stem per group and one shared reference line — the shape of Cédric Scherer's well-known "Evolution of a ggplot" student–teacher-ratio animation, redrawn in Qu.

It borrows the technique — jittered points per category, a bigger mean marker connected back to the axis by a short stem, one vertical line for the pooled average — not the original's palette or fonts. The colours come from palette(n), Qu's own hue-spaced categorical set, in keeping with how this project treats ggplot-inspired work generally: borrow the idea, keep Qu's own identity. The data is synthetic and shaped only to tell roughly the same story the original does (one region low and tight, another high and spread, one dramatic outlier) — not the real UNESCO figures the original chart used.

The whole program

# qu_beeswarm_continents.qu — a categorical dot plot with per-group means
# and a shared reference line, in the shape of Cedric Scherer's well-known
# "Evolution of a ggplot" student:teacher-ratio animation.
#
# Borrows the TECHNIQUE (jittered points per category, a bigger mean marker
# connected back by a short stem, one vertical line for the pooled average)
# without cloning the original's palette or fonts -- see
# book/src/stdlib/plotting.md's "Colours" section: `palette(n)` is Qu's own
# hue-spaced categorical set, used here instead of picking six colours by
# hand.
#
# The data is SYNTHETIC and illustrative only -- randn() draws per
# "continent", shaped to roughly the same story the original chart tells
# (Europe low and tight, Africa high and spread with one dramatic outlier)
# -- not the real UNESCO figures the original used. Swap `continent_data`
# for a real table and everything downstream is unchanged.
#
# Exercises:
#   randn(), uniform(), clip() — synthetic per-group samples and jitter
#   scatter(), plot()          — the swarm points and each mean's stem
#   yticks()/yticklabels()     — a categorical axis out of numeric y positions
#   vline(), text(), palette() — the shared reference line and its label

backend auto

#%% One row per "continent", bottom to top
names = ("Africa", "Oceania", "South America", "North America", "Asia", "Europe")
means  = (34,        20,       21,              16,              23,     14)
stds   = (11,         6,        5,               3,               8,      3)
counts = (32,        14,       15,              13,              30,     26)
colors = palette(len(names))

#%% Build each group's synthetic sample and the figure's pooled mean
function continent_data(n, m, s, seed)
    return clip(randn(n, seed = seed) * s + m, 3)
end function

group_samples = ()
for i in 0 to len(names) - 1
    xs = continent_data(counts[i], means[i], stds[i], 200 + i)
    if names[i] == "Africa"
        # One exaggerated outlier, echoing the original chart's single
        # extreme country without claiming to BE that country.
        xs = [xs, [86]]
    end if
    group_samples = append(group_samples, xs)
end for

# Pooled across every group -- built by concatenating Vec onto Vec, not
# starting from an empty `()` list, which `[a, b]` handles as generic list
# concatenation instead of a numeric hstack.
pooled = group_samples[0]
for i in 1 to len(group_samples) - 1
    pooled = [pooled, group_samples[i]]
end for
overall_avg = mean(pooled)

#%% Draw
figure()
figure_size(760, 480)
fontfamily("sans")

for i in 0 to len(names) - 1
    xs = group_samples[i]
    n = len(xs)
    ys = i + uniform(-0.16, 0.16, n, seed = 900 + i)
    scatter(xs, ys, color = colors[i], alpha = 0.45, ms = 5)
    m = mean(xs)
    plot([min(xs), m], [i, i], color = colors[i], width = 1.6)
    scatter([m], [i], color = colors[i], ms = 13, fill = true, edge = "#ffffff")
end for

vline(overall_avg, color = "#444444")
text(overall_avg + 1.5, len(names) - 0.4, "Worldwide average: {round(overall_avg, 1)}", color = "#444444")

xlim(0, 95)
ylim(-0.7, len(names) - 0.3)
yticks(0 to len(names) - 1)
yticklabels(names[0], names[1], names[2], names[3], names[4], names[5])
xlabel("Student-to-teacher ratio (synthetic data)")
title("The evolution of a beeswarm", align = "left")
grid off

savefig("catalog/qu_beeswarm_continents.svg")
print("wrote catalog/qu_beeswarm_continents.svg, worldwide average = {round(overall_avg, 1)}")

show plot

Source: catalog/qu_beeswarm_continents.qu