Start here
Which chart
You have a shape of data. Here is the preset, the plate, and the page that explains it.
You have a shape of data. The preset is the short spelling. The plate is that spelling, drawn. The link is the page that takes the lid off.
One series#
Numbers in order, against the index. The preset is line.
7.5 ┤
│
│
5.0 ┤
│
2.5 ┤
│
0.0 ┤
└┬─────────────┬─────────────┬─────────────┬
0 1 2 3
Plate 1. line over four values.
Linegrammar , and it prints the same bytes as the preset it grew from.
Categories#
A name and a number. The preset is bar. Bars rise from zero on a band axis.
7.5 ┤
│
│
5.0 ┤
│
2.5 ┤
│
0.0 ┤
└───────────────────────────────────────────
mon tue wed thu fri
Plate 2. bar over a week.
Sideways, stacked, and grouped are the bar mark , not new presets.
A distribution#
One column of measurements, and the shape it makes. The preset is hist. It chooses the bins. The count axis does not label a half.
body mass, g
100 ┤
│
75 ┤
│
│
50 ┤
│
25 ┤
│
0 ┤
└┬──────────┬──────────┬─────────┬──────────┬──────────┬
2000 3000 4000 5000 6000 7000
g
Plate 3. hist over 342 penguin body masses.
Bins, percent, and the cumulative climb are the statistics layer .
A density#
The same column, smoothed. The preset is density. The bandwidth is the knob.
kernel density, three bandwidths
──
½ Silverman
──
Silverman (default)
──
2× Silverman
600µ ┤
│
500µ ┤
400µ ┤
│
300µ ┤
│
200µ ┤
│
100µ ┤
0 ┤
└┬────────┬────────┬─────────┬────────┬─────────┬────────┬────────┬
1000 2000 3000 4000 5000 6000 7000 8000
body mass, g
Plate 4. One sample, three bandwidths. Silverman's rule is the middle one.
A box plot#
A number per category, and the spread. The preset is box_plot: type-7 quartiles, Tukey whiskers, outliers.
Range::over with body and marker
230 ┤
f │
l 220 ┤
i │ ━━━━━━━━
p 210 ┤
p │
e 200 ┤
r │
━━━━━━━━
, 190 ┤ ━━━━━━━━━
│
m 180 ┤
m │
170 ┤
└─────────────────────────────────────────────────────
Adelie Chinstrap Gentoo
Plate 5. The box plot, spelled with Range: whiskers, a body from q1 to q3, a marker at the median.
A violin#
The shape per category, instead of the five numbers. The preset is violin. It is two horizontal areas around a kernel density.
a violin is two horizontal areas
Area::horizontal
f 240 ┤
l │
i 230 ┤
p │
p │
e 220 ┤
r │
, │
210 ┤
m │
m 200 ┤
│
└┬────────┬───────┬────────┬────────┬────────┬───────┬
-0.06 -0.04 -0.02 0.00 0.02 0.04 0.06
Plate 6. One violin, spelled as two horizontal areas.
Two columns#
An x and a y. The preset is scatter.
penguin bills
d 22 ┤
e │
p 20 ┤
t │
h 18 ┤
, │
16 ┤
m │
m 14 ┤
│
└┬────────┬────────┬────────┬───────┬────────┬────────┬
30 35 40 45 50 55 60
length, mm
Plate 7. scatter over two columns.
Categories on a scatter#
The same two columns, and a name for each point. color_by paints the groups, names them in the legend, and cycles marker shapes when the output has no color.
flipper length against body mass
••
Adelie
••
Chinstrap
••
Gentoo
│
6000 ┤
│
│
5000 ┤
│
g │
│
4000 ┤
│
│
3000 ┤
│
└┬─────────┬──────────┬─────────┬─────────┬──────────┬─────────┬─
170 180 190 200 210 220 230
mm
Plate 8. Three species, one color_by channel.
The grammar builds this up one layer at a time.
A matrix#
Rows, columns, and a value in each cell. The preset is heatmap.
Cells::matrix — a value grid under a colormap
│
Jan ┤
35
Feb ┤
│
▓
30
Mar ┤
▓
│
▓
Apr ┤
▒
25
│
▒
May ┤
░
20
Jun ┤
░
│
░
└────────────────────────────────────────────────────── 15
00 04 08 12 16 20
Plate 9. Cells::matrix under a colormap. heatmap is the short spelling.
A logarithmic colormap, for weights that span decades, is scales .
Calendar time#
A timestamp and a number. The axis reads the span and labels hours, days, or years. You do not hand it a format string.
CO₂ at Mauna Loa (NOAA)
440 ┤
│
420 ┤
│
400 ┤
p │
p 380 ┤
m │
360 ┤
│
340 ┤
│
320 ┤
└─┬─────────┬────────┬─────────┬────────┬─────────┬────────┬──────
1960 1970 1980 1990 2000 2010 2020
Plate 10. Monthly CO₂ at Mauna Loa, 1958 onward, on a calendar axis.
A table first#
The summary you usually compute before you plot anything. The preset is describe.
flipper length, mm
│
Adelie ┤ 151 190.0 6.539 172 186 190 195 210
Chinstrap ┤ 68 195.8 7.132 178 191 196 201 212
Gentoo ┤ 123 217.2 6.485 203 212 216 221 231
│
└─────────────────────────────────────────────────────────────────────
count mean sd min p25 p50 p75 max
Plate 11. describe: count, mean, sd, min, quartiles, max. A gap is an em dash.
More points than pixels#
A long line. Past four points per column, the plot inserts M4 — first, last, minimum, and maximum in each column — and the pixels match drawing every point. A spike one sample wide stays.
200,000 points through M4
8 ┤
│
6 ┤
4 ┤
│
2 ┤
│
0 ┤
│
-2 ┤
-4 ┤
└┬────────┬─────────┬────────┬─────────┬────────┬────────┬─────────┬────────┬
0 25k 50k 75k 100k 125k 150k 175k 200k
Plate 12. 200,000 points. The three one-sample spikes survive.
The measurement is on the performance page.