Chart guide

Code on the page

Type ``` and Enter for a block of Ruby; Enter three times at its end to leave it. Blocks run top to bottom and share what they define, so data in one block can be charted in the next. A chart shows under its block a moment after you stop typing.

visits = dataset month: [ 1, 2, 3, 4, 5, 6 ], patients: [ 42, 51, 48, 63, 70, 74 ]

chart :line, visits, x: :month, y: :patients, title: "Patients seen"

visits = dataset month: [ 1, 2, 3, 4, 5, 6 ], patients: [ 42, 51, 48, 63, 70, 74 ]

chart :line, visits do
  title "Patients seen"
  x :month, label: "Month"
  y :patients, label: "Patients", from: 0
  color :red
  hline 60, label: "Capacity"
  caption "Patients seen in clinic each month."
end

Data

bp = dataset <<~CSV
  Age,Systolic BP,Arm
  34,120,drug
  51,135,placebo
  62,141,drug
CSV

bp = dataset age: [ 34, 51, 62 ], systolic: [ 120, 135, 141 ]
bp = dataset [ { age: 34, systolic: 120 }, { age: 51, systolic: 135 } ]

trials = dataset week: [ 0, 4, 8, 0, 4, 8 ], arm: %w[ drug drug drug placebo placebo placebo ],
  response: [ 0.1, 0.4, 0.6, 0.1, 0.2, 0.25 ]

drug = trials.where(arm: "drug")
later = trials.where { week >= 4 }
means = trials.summarize(by: :arm) { |rows| { mean: rows.mean(:response) } }
percent = trials.derive(:percent) { response * 100 }
puts means.to_a.inspect

Or drop a CSV, TSV or JSON file on the page: it's kept under Data, and a line reading it goes in, like trials = import "trials". A web address works too: import "https://…/cases.csv".

More verbs: sort_by, order, first, count_by, only, except, rename, relabel, join, longer, group_by, and mean, median, sd, quantile, sum, min, max of a column.

Kinds of chart

trials = dataset week: [ 0, 4, 8, 12, 0, 4, 8, 12 ], arm: %w[ drug drug drug drug placebo placebo placebo placebo ],
  response: [ 0.1, 0.35, 0.5, 0.6, 0.1, 0.15, 0.2, 0.22 ]

chart :line, trials, x: :week, y: :response, color: :arm, y_format: :percent

patients = dataset bmi: [ 21, 24, 27, 29, 31, 34, 36 ], sbp: [ 118, 121, 128, 131, 137, 142, 149 ]

chart :scatter, patients, x: :bmi, y: :sbp, trend: true

arms = dataset arm: %w[ Placebo Low High ], responders: [ 0.21, 0.38, 0.52 ]

chart :bar, arms, x: :arm, y: :responders, labels: true, y_format: :percent

beds = dataset week: [ 1, 2, 3, 1, 2, 3 ], ward: %w[ ICU ICU ICU General General General ], patients: [ 8, 12, 9, 30, 41, 37 ]

chart :area, beds, x: :week, y: :patients, color: :ward, stack: true

dose = dataset day: [ 0, 3, 7, 14 ], mg: [ 5, 10, 20, 20 ]

chart :step, dose, x: :day, y: :mg

ages = dataset age: [ 23, 31, 34, 38, 41, 44, 45, 47, 52, 53, 56, 58, 61, 64, 67, 71, 78 ]

chart :histogram, ages, x: :age, bins: 6

levels = dataset arm: %w[ A A A A A B B B B B ], crp: [ 3, 5, 6, 8, 21, 2, 3, 3, 4, 6 ]

chart :box, levels, x: :arm, y: :crp, points: true

studies = dataset [
  { study: "Smith 2019", ratio: 0.82, low: 0.61, high: 1.10, n: 120 },
  { study: "Lee 2021", ratio: 0.66, low: 0.51, high: 0.86, n: 300 },
  { study: "Overall", ratio: 0.72, low: 0.60, high: 0.87 }
]

chart :forest, studies, x: :ratio, label: :study, error: { low: :low, high: :high }, size: :n, log: :x

cohort = dataset months: [ 3, 5, 7, 9, 12, 14, 18, 2, 4, 6, 6, 8, 15, 1 ],
  died: [ 1, 1, 0, 1, 0, 1, 0, 1, 1, 1, 0, 1, 1, 1 ], arm: %w[ drug ] * 7 + %w[ placebo ] * 7

chart :survival, cohort, time: :months, event: :died, color: :arm

baseline = dataset arm: %w[ Drug Placebo ], n: [ 152, 149 ], age: [ 61.2, 60.8 ]

table baseline, title: "Baseline", caption: "Mean age, by arm."

Every setting

In a block as title "…", or as a keyword: title: "…".

Words

title

Above the chart, bold title "Patients seen"

subtitle

Under the title subtitle "Intention to treat"

caption

Under the chart, after “Figure 1.” caption "By month, 2024."

source

Smallest, at the foot source "Clinic records"

numbered

Leave out “Figure 1.” numbered false

Axes

x, y

Which column goes on each axis x :week

y [ … ]

A line or bar per column y [ :sbp, :dbp ]

label:

The axis title y :sbp, label: "Systolic (mmHg)"

format:

How numbers read y format: :percent

from:, to:

Where the axis starts and ends y from: 0, to: 1

ticks:

Roughly how many marks x ticks: 10

scale:

:linear, :log, :date or :band x scale: :log

log

Log scale on an axis log :x

x_label: …

Any axis setting, as a keyword chart :line, d, y_label: "Mean"

Colour

color

One colour for everything color :red

color :column

A colour per value of a column color :arm

color … =>

Choose the colours color :arm, "drug" => :red, "placebo" => :faint

opacity

How solid fills are, 0 to 1 opacity 0.5

Marks on the chart

error

Error bars, or a band on a line error low: :ci_low, high: :ci_high

hline

A line across, at a value hline 0.5, label: "Target"

vline

A line up, at a value vline 8, label: "Surgery"

band

A shaded span band x: 4..8, label: "Treatment"

note

Words at a point note 10, 0.7, "Plateau"

trend

Least-squares line (scatter) trend true

Look

legend

:top, :bottom, :right, :end or :none legend :bottom

aspect

Width over height aspect 2

grid

Grid lines: false, true or :both grid :both

points

Dots on lines; values on boxes points true

width

Line width width 2

dashed

Dashed lines dashed true

size

Point or square size, or a column size :n

stack

Stack bars or areas stack true

horizontal

Bars lying down horizontal true

labels

Values on the bars labels true

bins

Bars in a histogram bins 20

Survival and forest

time, event

Follow-up, and whether the event happened time :months; event :died

logrank

Leave out the log-rank p logrank false

label

The row names in a forest plot label :study

reference

Line of no effect: 1 on log, 0 otherwise reference false

Formats

:percent

0.25 → 25%

:integer

1234.6 → 1,235

:thousands

12500 → 12,500

:decimal

two places

:scientific

1.20e+04

"%.1f mg"

printf, any way you like

->(v) { … }

a lambda

Colours

• :ink
• :red
• :green
• :blue
• :yellow
• :purple
• :orange
• :teal
• :soft
• :faint

Or any "#rrggbb".