A graph is more than decoration. A graph is an argument, a claim about what the numbers mean, made in a language (position, length, color) that the eye reads faster than any sentence. And like any argument, a graph can be honest or it can mislead, sometimes without its maker even noticing. This chapter is about making the honest kind, and recognizing the other kind. It is also about what a graph can reveal: Anscombe’s quartet showed four datasets with identical statistics but four completely different shapes, a reminder that a picture often catches what a summary number hides.
The examples here come from a real project - The Sports Page, a daily newsletter (a sister project to this book) that takes one strange sports number and tells the truth about what it does and does not mean. Its house rule is the same one that runs through this whole book: describe what happened, predict what will happen, and never dress up one as the other. Good graphics are how you keep that promise to your reader.
The first decision is not color or font - it is geometry. What shape of mark answers your question? A change over time wants a line. Here is a true story: over the last four decades, major-league baseball became a strikeout sport. Watch it in one line:
Every figure here uses real data that ships with the book: data/mlb_league_by_year.csv, data/mlb_hitters_2026.csv, and data/nfl_draft_by_team.csv. Read them directly - no setup needed. See Getting the Book’s Data if you want the setup files for your language.
library(tidyverse) # dplyr, ggplot2, purrr, tibble, readr, stringr, forcats
source("_common.R") # book-wide helpers: round2(), fmt_p(), tidy2()
mlb <- read_csv("data/mlb_league_by_year.csv", show_col_types = FALSE)
ggplot(mlb, aes(year, K_9)) +
geom_line(color = "steelblue", linewidth = 1) +
geom_point(color = "steelblue", size = 1.4) +
labs(x = NULL, y = "Strikeouts per 9 innings",
title = "The strikeout revolution") +
theme_book()
* A trend over time wants a line.
GET DATA /TYPE=TXT /FILE='data/mlb_league_by_year.csv'
/DELIMITERS=',' /QUALIFIER='"' /FIRSTCASE=2
/VARIABLES=year F16.6 OPS F16.6 K_PA F16.6 BB_PA F16.6 HR_PA F16.6 AVG F16.6 ERA F16.6 K_9 F16.6 BB_9 F16.6 HR_9 F16.6.
GRAPH /LINE(SIMPLE) = VALUE(K_9) BY year./tmp/RtmpmOeUNM/file393652eac23a.sps:6.8-6.11: error: GRAPH: LINE is not yet
implemented.
6 | GRAPH /LINE(SIMPLE) = VALUE(K_9) BY year.
| ^~~~using CSV, DataFrames, Plots
mlb = CSV.read("data/mlb_league_by_year.csv", DataFrame)
plot(mlb.year, mlb.K_9, linewidth = 2, marker = :circle, legend = false,
ylabel = "Strikeouts per 9 innings", title = "The strikeout revolution")
import pandas as pd
import matplotlib.pyplot as plt
mlb = pd.read_csv("data/mlb_league_by_year.csv")
plt.plot(mlb.year, mlb.K_9, color="steelblue", linewidth=2, marker="o", ms=4)
plt.ylabel("Strikeouts per 9 innings")
plt.title("The strikeout revolution")
plt.show()[<matplotlib.lines.Line2D object at 0x7f50b073f4a0>]
Text(0, 0.5, 'Strikeouts per 9 innings')
Text(0.5, 1.0, 'The strikeout revolution')
A line says “these points are connected in sequence” - exactly the claim you want for a time trend. A bar chart here would be wrong: bars say “these are separate categories to compare,” which is a different (and misleading) sentence. The geometry is the argument. Trend → line; category comparison → bar; one distribution → histogram; two variables → scatter. Pick the mark that makes the claim you actually mean.
Same data, two axes, two very different impressions. Batting average has drifted down over the same period, from about .257 to about .242, a real but modest slide. First, watch a zoomed-in y-axis make that modest slide look like a catastrophe:
ggplot(mlb, aes(year, AVG)) +
geom_line(color = "firebrick", linewidth = 1) +
coord_cartesian(ylim = c(0.242, 0.258)) + # axis zoomed to the data
labs(x = NULL, y = "AVG", title = "“Hitting is DEAD”") +
theme_book()
GET DATA /TYPE=TXT /FILE='data/mlb_league_by_year.csv'
/DELIMITERS=',' /QUALIFIER='"' /FIRSTCASE=2
/VARIABLES=year F16.6 OPS F16.6 K_PA F16.6 BB_PA F16.6 HR_PA F16.6 AVG F16.6 ERA F16.6 K_9 F16.6 BB_9 F16.6 HR_9 F16.6.
* Dishonest axis: zoom the y scale tight and a 15-point drift looks
* like a cliff. In the chart editor you would set the y-axis minimum to
* .242 and maximum to .258; in syntax, GGRAPH carries the scale range.
GGRAPH
/GRAPHDATASET NAME="g" VARIABLES=year AVG
/GRAPHSPEC SOURCE=INLINE
/FRAME INNER=YES.
BEGIN GPL
SOURCE: s=userSource(id("g"))
DATA: year=col(source(s), name("year"))
DATA: AVG=col(source(s), name("AVG"))
SCALE: linear(dim(2), min(0.242), max(0.258))
GUIDE: text.title(label("Hitting is DEAD"))
ELEMENT: line(position(year*AVG))
END GPL./tmp/RtmpmOeUNM/file3936231f3185.sps:7.1-7.6: error: GGRAPH: GGRAPH is not yet
implemented.
7 | GGRAPH
| ^~~~~~
/tmp/RtmpmOeUNM/file3936231f3185.sps:11.1-11.9: error: Unknown command `BEGIN
GPL'.
11 | BEGIN GPL
| ^~~~~~~~~using CSV, DataFrames, Plots
mlb = CSV.read("data/mlb_league_by_year.csv", DataFrame)
plot(mlb.year, mlb.AVG, linewidth = 2, color = :firebrick, legend = false,
ylims = (0.242, 0.258), # axis zoomed to the data
ylabel = "AVG", title = "\"Hitting is DEAD\"")
import pandas as pd
import matplotlib.pyplot as plt
mlb = pd.read_csv("data/mlb_league_by_year.csv")
plt.plot(mlb.year, mlb.AVG, color="firebrick", linewidth=2)
plt.ylim(0.242, 0.258) # axis zoomed to the data
plt.ylabel("AVG"); plt.title('"Hitting is DEAD"')
plt.show()[<matplotlib.lines.Line2D object at 0x7f59d80b6150>]
(0.242, 0.258)
Text(0, 0.5, 'AVG')
Text(0.5, 1.0, '"Hitting is DEAD"')
Now the exact same numbers, on an axis anchored to a meaningful baseline:
ggplot(mlb, aes(year, AVG)) +
geom_line(color = "firebrick", linewidth = 1) +
coord_cartesian(ylim = c(0, 0.3)) + # axis from a meaningful zero
labs(x = NULL, y = "AVG", title = "“Hitting drifted down a bit”") +
theme_book()
GET DATA /TYPE=TXT /FILE='data/mlb_league_by_year.csv'
/DELIMITERS=',' /QUALIFIER='"' /FIRSTCASE=2
/VARIABLES=year F16.6 OPS F16.6 K_PA F16.6 BB_PA F16.6 HR_PA F16.6 AVG F16.6 ERA F16.6 K_9 F16.6 BB_9 F16.6 HR_9 F16.6.
* Honest axis: from a meaningful zero, the same drift is a gentle slope.
GGRAPH
/GRAPHDATASET NAME="g" VARIABLES=year AVG
/GRAPHSPEC SOURCE=INLINE
/FRAME INNER=YES.
BEGIN GPL
SOURCE: s=userSource(id("g"))
DATA: year=col(source(s), name("year"))
DATA: AVG=col(source(s), name("AVG"))
SCALE: linear(dim(2), min(0), max(0.3))
GUIDE: text.title(label("Hitting drifted down a bit"))
ELEMENT: line(position(year*AVG))
END GPL./tmp/RtmpmOeUNM/file393627e69ff6.sps:5.1-5.6: error: GGRAPH: GGRAPH is not yet
implemented.
5 | GGRAPH
| ^~~~~~
/tmp/RtmpmOeUNM/file393627e69ff6.sps:9.1-9.9: error: Unknown command `BEGIN
GPL'.
9 | BEGIN GPL
| ^~~~~~~~~using CSV, DataFrames, Plots
mlb = CSV.read("data/mlb_league_by_year.csv", DataFrame)
plot(mlb.year, mlb.AVG, linewidth = 2, color = :firebrick, legend = false,
ylims = (0, 0.3), # axis from a meaningful zero
ylabel = "AVG", title = "\"Hitting drifted down a bit\"")
import pandas as pd
import matplotlib.pyplot as plt
mlb = pd.read_csv("data/mlb_league_by_year.csv")
plt.plot(mlb.year, mlb.AVG, color="firebrick", linewidth=2)
plt.ylim(0, 0.3) # axis from a meaningful zero
plt.ylabel("AVG"); plt.title('"Hitting drifted down a bit"')
plt.show()[<matplotlib.lines.Line2D object at 0x7f22232b3380>]
(0.0, 0.3)
Text(0, 0.5, 'AVG')
Text(0.5, 1.0, '"Hitting drifted down a bit"')
Same numbers, opposite impressions. The first axis turns a 15-point drift into what looks like a cliff; the second, anchored to a meaningful baseline, shows the truth: hitting softened, it did not collapse. Neither axis is always right. A zoomed axis is legitimate when small differences genuinely matter (and you say so), but the choice is never neutral. Ask what the axis is doing to the size of the effect, and whether that is honest.
A truncated y-axis is the single most common way a well-meaning person makes a small effect look enormous (or the reverse). Before you publish a chart, look at your axis limits and ask: would this effect still look this big if I started from zero? If the honest version looks unremarkable, the honest answer might be that the effect is unremarkable.
You have three trends to show at once: strikeouts, walks, and home runs per plate appearance. The temptation is to put them all on one chart with two or three y-axes. It is better not to. Dual-axis charts let you manufacture almost any relationship by rescaling one axis, and readers cannot tell which line belongs to which scale. The clearer tool is small multiples: the same chart repeated, one panel per series, on comparable axes.
long <- mlb |>
select(year, Strikeouts = K_PA, Walks = BB_PA, `Home runs` = HR_PA) |>
pivot_longer(-year, names_to = "kind", values_to = "rate")
ggplot(long, aes(year, rate)) +
geom_line(color = "steelblue", linewidth = 0.9) +
facet_wrap(~ kind, scales = "free_y") + # one panel per series
labs(x = NULL, y = "per plate appearance") +
theme_book()
* Valid SPSS, shown rather than run: PSPP has implemented neither
* VARSTOCASES (which stacks the three rate columns into long form) nor
* GGRAPH's PANEL clause, which is how SPSS draws small multiples.
GET DATA /TYPE=TXT /FILE='data/mlb_league_by_year.csv'
/DELIMITERS=',' /QUALIFIER='"' /FIRSTCASE=2
/VARIABLES=year F16.6 K_PA F16.6 BB_PA F16.6 HR_PA F16.6.
VARSTOCASES /MAKE rate FROM K_PA BB_PA HR_PA /INDEX=kind.
GGRAPH /GRAPHDATASET NAME="g" VARIABLES=year rate kind
/GRAPHSPEC SOURCE=INLINE.
BEGIN GPL
SOURCE: s=userSource(id("g"))
DATA: year=col(source(s), name("year"))
DATA: rate=col(source(s), name("rate"))
DATA: kind=col(source(s), name("kind"), unit.category())
ELEMENT: line(position(year*rate*kind))
END GPL.using CSV, DataFrames, Plots
mlb = CSV.read("data/mlb_league_by_year.csv", DataFrame)
# one panel per series
plot([plot(mlb.year, mlb[!, c], title = String(c), legend = false)
for c in [:K_PA, :BB_PA, :HR_PA]]..., layout = (1, 3))
import pandas as pd
import matplotlib.pyplot as plt
mlb = pd.read_csv("data/mlb_league_by_year.csv")
long = mlb.melt(id_vars="year", value_vars=["K_PA", "BB_PA", "HR_PA"],
var_name="kind", value_name="rate")
fig, axes = plt.subplots(1, 3, figsize=(10, 3)) # one panel per series
for ax, (kind, g) in zip(axes, long.groupby("kind")):
ax.plot(g.year, g.rate, color="steelblue")
ax.set_title(kind)
axes[0].set_ylabel("per plate appearance")
fig.tight_layout(); plt.show()[<matplotlib.lines.Line2D object at 0x7ff9b01b3d70>]
Text(0.5, 1.0, 'BB_PA')
[<matplotlib.lines.Line2D object at 0x7ff9b01b3fb0>]
Text(0.5, 1.0, 'HR_PA')
[<matplotlib.lines.Line2D object at 0x7ff9b01d4290>]
Text(0.5, 1.0, 'K_PA')
Text(0, 0.5, 'per plate appearance')
Each panel tells its own honest story (strikeouts soared, walks held steady, homers crept up), and no rescaling trick can fake a relationship between them. When you have many series, repeat the chart; don’t overload it.
For comparing a quantity across categories, the eye reads position along a common scale far more accurately than it reads the angles of a pie. So a bar chart beats a pie chart almost always - and a sorted bar chart beats an unsorted one, because the ordering is itself information. Which NFL teams draft worst? Thirty-two teams is already far too many for a pie; rank them:
nfl <- read_csv("data/nfl_draft_by_team.csv", show_col_types = FALSE)
ggplot(nfl, aes(x = fct_reorder(team, rate_bust_a), y = rate_bust_a)) +
geom_col(fill = "grey35") +
coord_flip() + # horizontal: labels stay readable
labs(x = NULL, y = "Bust rate (no Pro Bowl)", title = "Draft busts by team") +
theme_book()
* Rank, don't pie: a sorted horizontal bar chart of bust rate by team.
GET DATA /TYPE=TXT /FILE='data/nfl_draft_by_team.csv'
/DELIMITERS=',' /QUALIFIER='"' /FIRSTCASE=2
/VARIABLES=team A24 n_picks F16.6 n_judged F16.6 n_hof F16.6 n_bust_a F16.6 n_bust_b F16.6 n_bust_c F16.6 n_triple_bust F16.6 n_clean_hit F16.6 rate_bust_a F16.6 rate_bust_b F16.6 rate_bust_c F16.6 rate_triple_bust F16.6 rate_clean_hit F16.6 hof_per_pick F16.6.
SORT CASES BY rate_bust_a (A).
* MEAN() rather than VALUE() here: each team has exactly one row, so the
* two are identical, and MEAN() is accepted by every SPSS-family engine.
GRAPH /BAR(SIMPLE) = MEAN(rate_bust_a) BY team.
* In the chart editor, transpose to horizontal so the team labels stay
* readable - the equivalent of coord_flip().
using CSV, DataFrames, Plots
nfl = CSV.read("data/nfl_draft_by_team.csv", DataFrame)
sort!(nfl, :rate_bust_a) # sorted: the ranking IS the message
bar(nfl.team, nfl.rate_bust_a, orientation = :horizontal, legend = false,
xlabel = "Bust rate (no Pro Bowl)")32×15 DataFrame
Row │ team n_picks n_judged n_hof n_bust_a n_bust_b n_bust_c n_triple_bust n_clean_hit rate_bust_a rate_bust_b rate_bust_c rate_triple_bust rate_clean_hit hof_per_pick
│ String3 Int64 Int64 Int64 Int64 Int64 Int64 Int64 Int64 Float64 Float64 Float64 Float64 Float64 Float64
─────┼────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
1 │ WAS 35 29 3 11 7 8 6 16 0.3793 0.2414 0.2759 0.2069 0.5517 0.0857
2 │ BAL 33 25 3 10 3 3 2 15 0.4 0.12 0.12 0.08 0.6 0.0909
3 │ KC 42 36 2 17 10 13 10 18 0.4722 0.2778 0.3611 0.2778 0.5 0.0476
4 │ MIN 48 40 3 19 8 12 8 20 0.475 0.2 0.3 0.2 0.5 0.0625
5 │ BUF 46 42 2 20 10 13 9 20 0.4762 0.2381 0.3095 0.2143 0.4762 0.0435
6 │ TEN 43 37 2 18 4 8 4 18 0.4865 0.1081 0.2162 0.1081 0.4865 0.0465
7 │ HOU 22 18 1 9 2 2 2 9 0.5 0.1111 0.1111 0.1111 0.5 0.0455
8 │ CAR 31 25 2 13 5 6 5 12 0.52 0.2 0.24 0.2 0.48 0.0645
9 │ DAL 42 36 4 19 8 11 8 17 0.5278 0.2222 0.3056 0.2222 0.4722 0.0952
10 │ LAC 43 36 2 19 6 10 6 16 0.5278 0.1667 0.2778 0.1667 0.4444 0.0465
11 │ ATL 48 41 1 22 9 14 8 17 0.5366 0.2195 0.3415 0.1951 0.4146 0.0208
12 │ ARI 39 33 1 18 9 12 9 14 0.5455 0.2727 0.3636 0.2727 0.4242 0.0256
13 │ NYJ 53 44 1 24 12 15 11 20 0.5455 0.2727 0.3409 0.25 0.4545 0.0189
14 │ CHI 44 39 2 22 9 11 9 17 0.5641 0.2308 0.2821 0.2308 0.4359 0.0455
15 │ IND 44 40 6 23 7 9 6 17 0.575 0.175 0.225 0.15 0.425 0.1364
16 │ PHI 43 36 0 21 12 13 11 14 0.5833 0.3333 0.3611 0.3056 0.3889 0.0
17 │ NO 46 39 1 23 7 11 6 15 0.5897 0.1795 0.2821 0.1538 0.3846 0.0217
18 │ DET 52 44 2 26 10 13 10 18 0.5909 0.2273 0.2955 0.2273 0.4091 0.0385
19 │ TB 42 37 2 22 7 9 4 14 0.5946 0.1892 0.2432 0.1081 0.3784 0.0476
20 │ NE 49 44 2 27 10 15 10 17 0.6136 0.2273 0.3409 0.2273 0.3864 0.0408
21 │ SEA 43 37 4 23 7 10 7 14 0.6216 0.1892 0.2703 0.1892 0.3784 0.093
22 │ SF 50 45 4 28 13 18 13 17 0.6222 0.2889 0.4 0.2889 0.3778 0.08
23 │ PIT 46 41 3 26 10 14 9 14 0.6341 0.2439 0.3415 0.2195 0.3415 0.0652
24 │ CLE 42 39 1 25 13 16 11 13 0.641 0.3333 0.4103 0.2821 0.3333 0.0238
25 │ DEN 38 34 1 22 7 13 7 12 0.6471 0.2059 0.3824 0.2059 0.3529 0.0263
26 │ MIA 44 37 1 24 12 15 12 13 0.6486 0.3243 0.4054 0.3243 0.3514 0.0227
27 │ JAX 36 27 1 18 5 9 5 9 0.6667 0.1852 0.3333 0.1852 0.3333 0.0278
28 │ LAR 51 50 3 34 12 14 10 15 0.68 0.24 0.28 0.2 0.3 0.0588
29 │ NYG 49 41 1 28 9 10 9 13 0.6829 0.2195 0.2439 0.2195 0.3171 0.0204
30 │ GB 49 42 1 29 13 15 10 10 0.6905 0.3095 0.3571 0.2381 0.2381 0.0204
31 │ CIN 52 46 1 33 17 18 17 12 0.7174 0.3696 0.3913 0.3696 0.2609 0.0192
32 │ LV 44 44 3 32 12 16 10 11 0.7273 0.2727 0.3636 0.2273 0.25 0.0682
import pandas as pd
import matplotlib.pyplot as plt
nfl = pd.read_csv("data/nfl_draft_by_team.csv")
nfl = nfl.sort_values("rate_bust_a") # sorted: the ranking IS the message
plt.figure(figsize=(6, 8))
plt.barh(nfl.team, nfl.rate_bust_a, color="0.35")
plt.xlabel("Bust rate (no Pro Bowl)")
plt.tight_layout(); plt.show()<Figure size 1152x1536 with 0 Axes>
<BarContainer object of 32 artists>
Text(0.5, 0, 'Bust rate (no Pro Bowl)')
Sorted, horizontal, one bar per team: you can read off the best and worst drafters quickly, and every team is comparable on the same axis. A pie chart of these same 32 numbers would be very hard to read; you could not tell 22% from 25% by eye, and the ranking would disappear. Pie charts work for maybe two or three slices; past that, rank a bar chart.
A single summary number - a mean, a bar - throws away the shape of the data, and the shape is usually the story. Do not draw a bar of averages when you could draw the whole distribution. Here are this season’s qualified hitters; a histogram shows the spread of performance a mean would erase:
h <- read_csv("data/mlb_hitters_2026.csv", show_col_types = FALSE) |>
filter(PA >= 50)
ggplot(h, aes(OPS)) +
geom_histogram(aes(y = after_stat(density)), bins = 30,
fill = "steelblue", color = "white") +
# a normal with this data's own mean and SD - here the curve is a REFERENCE,
# not a claim: the gap between bars and curve is the interesting part
stat_function(fun = dnorm, args = list(mean = mean(h$OPS), sd = sd(h$OPS)),
colour = "grey30", linewidth = 0.9, linetype = "dashed") +
geom_vline(xintercept = median(h$OPS), color = "firebrick", linewidth = 1) +
labs(x = "OPS", y = "density", title = "Most hitters are ordinary; a few are stars") +
theme_book()
* Show the distribution, not just the mean.
GET DATA /TYPE=TXT /FILE='data/mlb_hitters_2026.csv'
/DELIMITERS=',' /QUALIFIER='"' /FIRSTCASE=2
/VARIABLES=pid F16.6 name A24 team A24 PA F16.6 AB F16.6 H F16.6 HR F16.6 BB F16.6 SO F16.6 AVG F16.6 OBP F16.6 SLG F16.6 OPS F16.6.
SELECT IF (PA >= 50).
EXECUTE.
FREQUENCIES VARIABLES=OPS /STATISTICS=MEDIAN /HISTOGRAM. Statistics
+---------+---+
| |OPS|
+---------+---+
|N Valid |414|
| Missing| 0|
+---------+---+
|Median |.69|
+---------+---+
OPS
+--------------+---------+-------+-------------+------------------+
| |Frequency|Percent|Valid Percent|Cumulative Percent|
+--------------+---------+-------+-------------+------------------+
|Valid .200000 | 1| .2%| .2%| .2%|
| .276000 | 1| .2%| .2%| .5%|
| .332000 | 1| .2%| .2%| .7%|
| .377000 | 1| .2%| .2%| 1.0%|
| .410000 | 1| .2%| .2%| 1.2%|
| .417000 | 1| .2%| .2%| 1.4%|
| .420000 | 1| .2%| .2%| 1.7%|
| .425000 | 1| .2%| .2%| 1.9%|
| .435000 | 1| .2%| .2%| 2.2%|
| .438000 | 1| .2%| .2%| 2.4%|
| .445000 | 1| .2%| .2%| 2.7%|
| .465000 | 1| .2%| .2%| 2.9%|
| .468000 | 2| .5%| .5%| 3.4%|
| .470000 | 1| .2%| .2%| 3.6%|
| .471000 | 1| .2%| .2%| 3.9%|
| .481000 | 1| .2%| .2%| 4.1%|
| .482000 | 1| .2%| .2%| 4.3%|
| .490000 | 1| .2%| .2%| 4.6%|
| .494000 | 1| .2%| .2%| 4.8%|
| .498000 | 2| .5%| .5%| 5.3%|
| .505000 | 1| .2%| .2%| 5.6%|
| .515000 | 2| .5%| .5%| 6.0%|
| .516000 | 2| .5%| .5%| 6.5%|
| .519000 | 1| .2%| .2%| 6.8%|
| .520000 | 1| .2%| .2%| 7.0%|
| .523000 | 1| .2%| .2%| 7.2%|
| .526000 | 1| .2%| .2%| 7.5%|
| .530000 | 1| .2%| .2%| 7.7%|
| .532000 | 3| .7%| .7%| 8.5%|
| .533000 | 1| .2%| .2%| 8.7%|
| .534000 | 2| .5%| .5%| 9.2%|
| .535000 | 1| .2%| .2%| 9.4%|
| .537000 | 1| .2%| .2%| 9.7%|
| .538000 | 2| .5%| .5%| 10.1%|
| .542000 | 1| .2%| .2%| 10.4%|
| .546000 | 1| .2%| .2%| 10.6%|
| .547000 | 1| .2%| .2%| 10.9%|
| .549000 | 1| .2%| .2%| 11.1%|
| .551000 | 1| .2%| .2%| 11.4%|
| .554000 | 1| .2%| .2%| 11.6%|
| .555000 | 2| .5%| .5%| 12.1%|
| .556000 | 1| .2%| .2%| 12.3%|
| .560000 | 4| 1.0%| 1.0%| 13.3%|
| .561000 | 1| .2%| .2%| 13.5%|
| .562000 | 1| .2%| .2%| 13.8%|
| .563000 | 2| .5%| .5%| 14.3%|
| .565000 | 1| .2%| .2%| 14.5%|
| .568000 | 1| .2%| .2%| 14.7%|
| .569000 | 2| .5%| .5%| 15.2%|
| .570000 | 1| .2%| .2%| 15.5%|
| .573000 | 1| .2%| .2%| 15.7%|
| .574000 | 1| .2%| .2%| 15.9%|
| .578000 | 1| .2%| .2%| 16.2%|
| .583000 | 1| .2%| .2%| 16.4%|
| .584000 | 1| .2%| .2%| 16.7%|
| .585000 | 1| .2%| .2%| 16.9%|
| .586000 | 1| .2%| .2%| 17.1%|
| .587000 | 1| .2%| .2%| 17.4%|
| .588000 | 1| .2%| .2%| 17.6%|
| .589000 | 1| .2%| .2%| 17.9%|
| .591000 | 2| .5%| .5%| 18.4%|
| .593000 | 2| .5%| .5%| 18.8%|
| .595000 | 3| .7%| .7%| 19.6%|
| .597000 | 1| .2%| .2%| 19.8%|
| .599000 | 1| .2%| .2%| 20.0%|
| .600000 | 1| .2%| .2%| 20.3%|
| .601000 | 2| .5%| .5%| 20.8%|
| .602000 | 2| .5%| .5%| 21.3%|
| .603000 | 1| .2%| .2%| 21.5%|
| .605000 | 1| .2%| .2%| 21.7%|
| .606000 | 1| .2%| .2%| 22.0%|
| .607000 | 1| .2%| .2%| 22.2%|
| .608000 | 2| .5%| .5%| 22.7%|
| .609000 | 1| .2%| .2%| 22.9%|
| .611000 | 3| .7%| .7%| 23.7%|
| .612000 | 3| .7%| .7%| 24.4%|
| .613000 | 2| .5%| .5%| 24.9%|
| .616000 | 1| .2%| .2%| 25.1%|
| .618000 | 1| .2%| .2%| 25.4%|
| .619000 | 1| .2%| .2%| 25.6%|
| .622000 | 2| .5%| .5%| 26.1%|
| .623000 | 3| .7%| .7%| 26.8%|
| .624000 | 2| .5%| .5%| 27.3%|
| .626000 | 2| .5%| .5%| 27.8%|
| .627000 | 2| .5%| .5%| 28.3%|
| .629000 | 1| .2%| .2%| 28.5%|
| .630000 | 1| .2%| .2%| 28.7%|
| .631000 | 1| .2%| .2%| 29.0%|
| .634000 | 3| .7%| .7%| 29.7%|
| .635000 | 3| .7%| .7%| 30.4%|
| .637000 | 2| .5%| .5%| 30.9%|
| .638000 | 1| .2%| .2%| 31.2%|
| .641000 | 2| .5%| .5%| 31.6%|
| .643000 | 1| .2%| .2%| 31.9%|
| .644000 | 1| .2%| .2%| 32.1%|
| .645000 | 1| .2%| .2%| 32.4%|
| .646000 | 4| 1.0%| 1.0%| 33.3%|
| .647000 | 1| .2%| .2%| 33.6%|
| .649000 | 2| .5%| .5%| 34.1%|
| .650000 | 3| .7%| .7%| 34.8%|
| .651000 | 2| .5%| .5%| 35.3%|
| .652000 | 2| .5%| .5%| 35.7%|
| .653000 | 1| .2%| .2%| 36.0%|
| .654000 | 1| .2%| .2%| 36.2%|
| .656000 | 2| .5%| .5%| 36.7%|
| .657000 | 1| .2%| .2%| 37.0%|
| .659000 | 2| .5%| .5%| 37.4%|
| .660000 | 3| .7%| .7%| 38.2%|
| .661000 | 2| .5%| .5%| 38.6%|
| .662000 | 2| .5%| .5%| 39.1%|
| .663000 | 3| .7%| .7%| 39.9%|
| .664000 | 3| .7%| .7%| 40.6%|
| .665000 | 3| .7%| .7%| 41.3%|
| .666000 | 2| .5%| .5%| 41.8%|
| .667000 | 3| .7%| .7%| 42.5%|
| .668000 | 1| .2%| .2%| 42.8%|
| .669000 | 3| .7%| .7%| 43.5%|
| .670000 | 2| .5%| .5%| 44.0%|
| .671000 | 1| .2%| .2%| 44.2%|
| .672000 | 2| .5%| .5%| 44.7%|
| .674000 | 1| .2%| .2%| 44.9%|
| .675000 | 1| .2%| .2%| 45.2%|
| .676000 | 1| .2%| .2%| 45.4%|
| .677000 | 1| .2%| .2%| 45.7%|
| .678000 | 1| .2%| .2%| 45.9%|
| .681000 | 1| .2%| .2%| 46.1%|
| .682000 | 3| .7%| .7%| 46.9%|
| .684000 | 1| .2%| .2%| 47.1%|
| .685000 | 1| .2%| .2%| 47.3%|
| .686000 | 3| .7%| .7%| 48.1%|
| .687000 | 1| .2%| .2%| 48.3%|
| .688000 | 1| .2%| .2%| 48.6%|
| .689000 | 1| .2%| .2%| 48.8%|
| .690000 | 2| .5%| .5%| 49.3%|
| .691000 | 1| .2%| .2%| 49.5%|
| .693000 | 1| .2%| .2%| 49.8%|
| .694000 | 1| .2%| .2%| 50.0%|
| .695000 | 2| .5%| .5%| 50.5%|
| .696000 | 1| .2%| .2%| 50.7%|
| .697000 | 1| .2%| .2%| 51.0%|
| .698000 | 1| .2%| .2%| 51.2%|
| .699000 | 1| .2%| .2%| 51.4%|
| .700000 | 2| .5%| .5%| 51.9%|
| .702000 | 2| .5%| .5%| 52.4%|
| .703000 | 3| .7%| .7%| 53.1%|
| .704000 | 2| .5%| .5%| 53.6%|
| .706000 | 1| .2%| .2%| 53.9%|
| .707000 | 1| .2%| .2%| 54.1%|
| .708000 | 2| .5%| .5%| 54.6%|
| .709000 | 1| .2%| .2%| 54.8%|
| .710000 | 3| .7%| .7%| 55.6%|
| .711000 | 1| .2%| .2%| 55.8%|
| .713000 | 1| .2%| .2%| 56.0%|
| .714000 | 1| .2%| .2%| 56.3%|
| .715000 | 1| .2%| .2%| 56.5%|
| .716000 | 2| .5%| .5%| 57.0%|
| .717000 | 1| .2%| .2%| 57.2%|
| .720000 | 3| .7%| .7%| 58.0%|
| .721000 | 2| .5%| .5%| 58.5%|
| .722000 | 2| .5%| .5%| 58.9%|
| .723000 | 2| .5%| .5%| 59.4%|
| .724000 | 1| .2%| .2%| 59.7%|
| .725000 | 1| .2%| .2%| 59.9%|
| .726000 | 1| .2%| .2%| 60.1%|
| .729000 | 2| .5%| .5%| 60.6%|
| .730000 | 1| .2%| .2%| 60.9%|
| .731000 | 2| .5%| .5%| 61.4%|
| .732000 | 1| .2%| .2%| 61.6%|
| .733000 | 2| .5%| .5%| 62.1%|
| .734000 | 1| .2%| .2%| 62.3%|
| .735000 | 1| .2%| .2%| 62.6%|
| .736000 | 2| .5%| .5%| 63.0%|
| .738000 | 2| .5%| .5%| 63.5%|
| .739000 | 2| .5%| .5%| 64.0%|
| .740000 | 1| .2%| .2%| 64.3%|
| .741000 | 1| .2%| .2%| 64.5%|
| .743000 | 2| .5%| .5%| 65.0%|
| .744000 | 1| .2%| .2%| 65.2%|
| .745000 | 1| .2%| .2%| 65.5%|
| .746000 | 1| .2%| .2%| 65.7%|
| .747000 | 2| .5%| .5%| 66.2%|
| .748000 | 1| .2%| .2%| 66.4%|
| .749000 | 3| .7%| .7%| 67.1%|
| .751000 | 1| .2%| .2%| 67.4%|
| .752000 | 1| .2%| .2%| 67.6%|
| .753000 | 1| .2%| .2%| 67.9%|
| .754000 | 1| .2%| .2%| 68.1%|
| .755000 | 2| .5%| .5%| 68.6%|
| .757000 | 1| .2%| .2%| 68.8%|
| .760000 | 1| .2%| .2%| 69.1%|
| .762000 | 2| .5%| .5%| 69.6%|
| .763000 | 1| .2%| .2%| 69.8%|
| .765000 | 1| .2%| .2%| 70.0%|
| .766000 | 1| .2%| .2%| 70.3%|
| .767000 | 1| .2%| .2%| 70.5%|
| .768000 | 2| .5%| .5%| 71.0%|
| .770000 | 2| .5%| .5%| 71.5%|
| .771000 | 2| .5%| .5%| 72.0%|
| .772000 | 3| .7%| .7%| 72.7%|
| .773000 | 1| .2%| .2%| 72.9%|
| .774000 | 2| .5%| .5%| 73.4%|
| .775000 | 1| .2%| .2%| 73.7%|
| .778000 | 1| .2%| .2%| 73.9%|
| .779000 | 2| .5%| .5%| 74.4%|
| .781000 | 1| .2%| .2%| 74.6%|
| .782000 | 1| .2%| .2%| 74.9%|
| .784000 | 1| .2%| .2%| 75.1%|
| .786000 | 1| .2%| .2%| 75.4%|
| .787000 | 4| 1.0%| 1.0%| 76.3%|
| .789000 | 2| .5%| .5%| 76.8%|
| .790000 | 1| .2%| .2%| 77.1%|
| .791000 | 1| .2%| .2%| 77.3%|
| .792000 | 2| .5%| .5%| 77.8%|
| .793000 | 1| .2%| .2%| 78.0%|
| .794000 | 1| .2%| .2%| 78.3%|
| .795000 | 2| .5%| .5%| 78.7%|
| .796000 | 1| .2%| .2%| 79.0%|
| .797000 | 1| .2%| .2%| 79.2%|
| .798000 | 2| .5%| .5%| 79.7%|
| .801000 | 3| .7%| .7%| 80.4%|
| .802000 | 3| .7%| .7%| 81.2%|
| .803000 | 2| .5%| .5%| 81.6%|
| .804000 | 1| .2%| .2%| 81.9%|
| .807000 | 2| .5%| .5%| 82.4%|
| .809000 | 1| .2%| .2%| 82.6%|
| .810000 | 1| .2%| .2%| 82.9%|
| .817000 | 1| .2%| .2%| 83.1%|
| .819000 | 1| .2%| .2%| 83.3%|
| .820000 | 1| .2%| .2%| 83.6%|
| .822000 | 1| .2%| .2%| 83.8%|
| .823000 | 1| .2%| .2%| 84.1%|
| .824000 | 2| .5%| .5%| 84.5%|
| .825000 | 1| .2%| .2%| 84.8%|
| .828000 | 1| .2%| .2%| 85.0%|
| .832000 | 2| .5%| .5%| 85.5%|
| .833000 | 1| .2%| .2%| 85.7%|
| .834000 | 1| .2%| .2%| 86.0%|
| .835000 | 1| .2%| .2%| 86.2%|
| .838000 | 2| .5%| .5%| 86.7%|
| .841000 | 1| .2%| .2%| 87.0%|
| .842000 | 2| .5%| .5%| 87.4%|
| .843000 | 1| .2%| .2%| 87.7%|
| .844000 | 4| 1.0%| 1.0%| 88.6%|
| .845000 | 3| .7%| .7%| 89.4%|
| .850000 | 1| .2%| .2%| 89.6%|
| .851000 | 1| .2%| .2%| 89.9%|
| .853000 | 1| .2%| .2%| 90.1%|
| .854000 | 1| .2%| .2%| 90.3%|
| .855000 | 2| .5%| .5%| 90.8%|
| .856000 | 1| .2%| .2%| 91.1%|
| .861000 | 1| .2%| .2%| 91.3%|
| .863000 | 1| .2%| .2%| 91.5%|
| .864000 | 1| .2%| .2%| 91.8%|
| .865000 | 1| .2%| .2%| 92.0%|
| .878000 | 2| .5%| .5%| 92.5%|
| .882000 | 1| .2%| .2%| 92.8%|
| .883000 | 1| .2%| .2%| 93.0%|
| .884000 | 1| .2%| .2%| 93.2%|
| .895000 | 1| .2%| .2%| 93.5%|
| .898000 | 1| .2%| .2%| 93.7%|
| .900000 | 1| .2%| .2%| 94.0%|
| .905000 | 2| .5%| .5%| 94.4%|
| .908000 | 1| .2%| .2%| 94.7%|
| .909000 | 1| .2%| .2%| 94.9%|
| .912000 | 1| .2%| .2%| 95.2%|
| .913000 | 1| .2%| .2%| 95.4%|
| .914000 | 1| .2%| .2%| 95.7%|
| .915000 | 1| .2%| .2%| 95.9%|
| .926000 | 1| .2%| .2%| 96.1%|
| .930000 | 2| .5%| .5%| 96.6%|
| .931000 | 1| .2%| .2%| 96.9%|
| .932000 | 1| .2%| .2%| 97.1%|
| .935000 | 1| .2%| .2%| 97.3%|
| .938000 | 1| .2%| .2%| 97.6%|
| .942000 | 1| .2%| .2%| 97.8%|
| .949000 | 1| .2%| .2%| 98.1%|
| .953000 | 1| .2%| .2%| 98.3%|
| .958000 | 1| .2%| .2%| 98.6%|
| .959000 | 1| .2%| .2%| 98.8%|
| .964000 | 1| .2%| .2%| 99.0%|
| .976000 | 1| .2%| .2%| 99.3%|
| .981000 | 1| .2%| .2%| 99.5%|
| .998000 | 1| .2%| .2%| 99.8%|
| 1.092000| 1| .2%| .2%| 100.0%|
+--------------+---------+-------+-------------+------------------+
|Total | 414| 100.0%| | |
+--------------+---------+-------+-------------+------------------+using CSV, DataFrames, Statistics, Plots
h = filter(:PA => >=(50), CSV.read("data/mlb_hitters_2026.csv", DataFrame))
histogram(h.OPS, bins = 30, legend = false, xlabel = "OPS", ylabel = "hitters",
title = "Most hitters are ordinary; a few are stars")
vline!([median(h.OPS)], color = :firebrick, linewidth = 2)
import pandas as pd
import matplotlib.pyplot as plt
h = pd.read_csv("data/mlb_hitters_2026.csv").query("PA >= 50")
plt.hist(h.OPS, bins=30, color="steelblue", edgecolor="white")
plt.axvline(h.OPS.median(), color="firebrick", linewidth=2)
plt.xlabel("OPS"); plt.ylabel("hitters")
plt.title("Most hitters are ordinary; a few are stars")
plt.show()(array([ 1., 0., 1., 0., 1., 1., 0., 5., 3., 8., 11., 20., 20.,
33., 34., 49., 35., 37., 30., 35., 26., 24., 11., 10., 10., 5.,
3., 0., 0., 1.]), array([0.2 , 0.22973333, 0.25946667, 0.2892 , 0.31893333,
0.34866667, 0.3784 , 0.40813333, 0.43786667, 0.4676 ,
0.49733333, 0.52706667, 0.5568 , 0.58653333, 0.61626667,
0.646 , 0.67573333, 0.70546667, 0.7352 , 0.76493333,
0.79466667, 0.8244 , 0.85413333, 0.88386667, 0.9136 ,
0.94333333, 0.97306667, 1.0028 , 1.03253333, 1.06226667,
1.092 ]), <BarContainer object of 30 artists>)
<matplotlib.lines.Line2D object at 0x7f22fc2f2300>
Text(0.5, 0, 'OPS')
Text(0, 0.5, 'hitters')
Text(0.5, 1.0, 'Most hitters are ordinary; a few are stars')
And when you plot two variables against each other with hundreds of points, they pile up and hide their own density. The fix is transparency (or jitter): let the dark regions show where the data actually concentrate.
ggplot(h, aes(OBP, SLG)) +
geom_point(alpha = 0.35, color = "steelblue") + # alpha tames overplotting
labs(x = "On-base percentage", y = "Slugging percentage") +
theme_book()
GET DATA /TYPE=TXT /FILE='data/mlb_hitters_2026.csv'
/DELIMITERS=',' /QUALIFIER='"' /FIRSTCASE=2
/VARIABLES=pid F16.6 name A24 team A24 PA F16.6 AB F16.6 H F16.6 HR F16.6 BB F16.6 SO F16.6 AVG F16.6 OBP F16.6 SLG F16.6 OPS F16.6.
* Transparency tames overplotting. SPSS sets marker transparency in the
* chart editor, or through GGRAPH's transparency() aesthetic.
GGRAPH
/GRAPHDATASET NAME="g" VARIABLES=OBP SLG
/GRAPHSPEC SOURCE=INLINE.
BEGIN GPL
SOURCE: s=userSource(id("g"))
DATA: OBP=col(source(s), name("OBP"))
DATA: SLG=col(source(s), name("SLG"))
GUIDE: axis(dim(1), label("On-base percentage"))
GUIDE: axis(dim(2), label("Slugging percentage"))
ELEMENT: point(position(OBP*SLG), transparency.exterior(transparency."0.65"))
END GPL./tmp/RtmpmOeUNM/file393619bd9d7.sps:6.1-6.6: error: GGRAPH: GGRAPH is not yet
implemented.
6 | GGRAPH
| ^~~~~~
/tmp/RtmpmOeUNM/file393619bd9d7.sps:9.1-9.9: error: Unknown command `BEGIN
GPL'.
9 | BEGIN GPL
| ^~~~~~~~~using CSV, DataFrames, Statistics, Plots
h = CSV.read("data/mlb_hitters_2026.csv", DataFrame)
scatter(h.OBP, h.SLG, alpha = 0.35, legend = false, # alpha tames overplotting
xlabel = "On-base percentage", ylabel = "Slugging percentage")
import pandas as pd
import matplotlib.pyplot as plt
h = pd.read_csv("data/mlb_hitters_2026.csv")
plt.scatter(h.OBP, h.SLG, alpha=0.35, color="steelblue") # tames overplotting
plt.xlabel("On-base percentage"); plt.ylabel("Slugging percentage")
plt.show()<matplotlib.collections.PathCollection object at 0x7f56aa2478f0>
Text(0.5, 0, 'On-base percentage')
Text(0, 0.5, 'Slugging percentage')
The cloud’s shape, with its positive drift, its dense middle, and a few outliers in the upper right, is information a table of two means could never carry. That is much of the reason graphics exist.
Before a chart leaves your hands, run it past four questions:
A graph that survives all four is an argument you can stand behind.
Graphics are half of how we show data; the other half is the table, and tables have their own quiet craft, their own ways of telling the truth or hiding it. That is the next chapter. Then, with tables and graphics both in hand, we turn to the payoff of this whole book: watching these tools used on real published work, with every judgment call laid bare.