Skip to examples

A visual reference · R / ggplot2

R / ggplot2,
one change at a time

Complete source. One deliberate change. Its actual result. Follow a small field archive through measurements, distributions and publication graphics. Every chart starts from explicit local data and shows how an aesthetic mapping becomes a native figure.

The Org file is canonical. Read the searchable HTML beside your editor or use the PDF offline. All measurements and operational quantities are synthetic teaching fixtures.

Last changes: 2026-11-16. Original examples, pinned native rendering and portable source/configuration downloads.

Download PDFRendered tour · read offline
96worked examples
16topic families
1canonical Org source

Reading the tour

Three ways to see a change

Read the source

Each example is a complete file. Identify the syntax introduced before changing it.

Inspect the output

Open a gallery thumbnail, compare the native figure and follow a sibling example.

Make a change

Download the source/configuration ZIP, make the suggested edit and repeat its exact command.

Browse this tour with the gallery, search and topic filters. Stable IDs identify chapters and examples; Back to diagram overview returns to the corresponding thumbnail. The shared configuration sets appearance, never example data.

Reproduce the diagrams

The reviewed native toolchain is R 4.6.2, ggplot2 4.0.4 and Poppler 26.11.0. Exact checksums and publisher versions are recorded in VERSIONS.json. Install R 4.6.2, Poppler and a C/C++ compiler. Run python3 tools/install-R once to install the pinned package sources from tools/vendor/ into the ignored .R-library/ directory. This installation reads its local package index and needs no network. The renderer uses the bundled Source Sans 3 font and native Cairo PDF; Poppler creates SVG and PNG companions.

Install the shared publisher with make setup in its sibling directory. Activate that Python environment before using a standalone bundle, or install its pinned requirements.txt in a fresh environment. The published HTML embeds figures and works offline; keep the companion directories for individual downloads.

make full
make check
tools/render --version
tools/render examples/GGP-01-001.R
Command or configuration Meaning in this tour
tools/render SOURCE Run the pinned native pipeline and write PDF, SVG and PNG.
config/render.json Reviewed tool versions and bundled font checksums.
SOURCE_DATE_EPOCH=0 Fix native build-date inputs; the driver removes PDF date metadata.
make full Explicitly execute reviewed programs, then build HTML and PDF.
make Reuse verified local results; never execute examples implicitly.
make check Verify output hashes, example IDs and offline links.
Rscript --vanilla Evaluate the complete program without user startup files.
config/style.R Set the common theme, local font, warning policy and six-by-four-inch Cairo PDF export.
config/packages.json Exact package versions and checksums of bundled source archives.
tools/install-R Compile the reviewed package sources from a local file repository.
pdftocairo -svg Convert the native Cairo PDF to a self-contained SVG companion.

Download the complete reproduction bundle to run all examples, or inspect the shared render configuration. After installing system tools and the shared Python dependencies, preparation and rendering use no network. SVG is self-contained; native PDF pages are placed directly in the guide. Most layers are vector artwork; the raster example deliberately embeds its image layer. Repair sources rather than editing generated figures.

01 · Build a plot

Start with explicit local values, map the columns, and add one drawing layer.

GGP-01-001

Place the measurements

A point layer maps two numeric columns to positions. Use it to inspect paired measurements before adding a trend or statistical model.

Back to diagram overview

Compare with gnuplot plotting tour.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, temperature=c(12,14,13,16,15))
p <- ggplot(d, aes(day, temperature)) + geom_point()
export_plot(p, "GGP-01-001")
ParameterMeaning in this example
aes(day, temperature)Map the two columns to horizontal and vertical positions.
geom_point()Draw one point for every observation.

Render this diagram

tools/render examples/GGP-01-001.R

Try changing: Change the fourth temperature to 18.

Actual output

Place the measurements
GGP-01-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-01-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-01-002

Map a third variable to color

An aesthetic mapping can carry a group identity as well as a position. Use a color legend when the field stations need to remain distinguishable.

Back to diagram overview

Compare with GGP-01-001 · Place the measurements.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=4),
                day=rep(1:4,2), reading=c(8,12,11,15,13,10,14,12))
p <- ggplot(d, aes(day, reading, colour=station)) + geom_point(size=3)
export_plot(p, "GGP-01-002")
ParameterMeaning in this example
colour=stationMap station identity to a color scale and legend.

Render this diagram

tools/render examples/GGP-01-002.R

Try changing: Rename the two stations.

Actual output

Map a third variable to color
GGP-01-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-01-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-01-003

Set a constant appearance

A constant outside aes applies to every mark. Use it for a presentation choice that does not represent another data variable.

Back to diagram overview

Compare with GGP-01-002 · Map a third variable to color.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, temperature=c(12,14,13,16,15))
p <- ggplot(d, aes(day, temperature)) +
  geom_point(colour="#00549F", size=3)
export_plot(p, "GGP-01-003")
ParameterMeaning in this example
colour outside aesSet a common color rather than mapping a data field.

Render this diagram

tools/render examples/GGP-01-003.R

Try changing: Change the fixed color to #C45A18.

Actual output

Set a constant appearance
GGP-01-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-01-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-01-004

Connect an ordered series

A line layer connects values in horizontal order. Use it for readings whose day coordinate defines a meaningful sequence.

Back to diagram overview

Compare with GGP-01-003 · Set a constant appearance.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, temperature=c(12,14,13,16,15))
p <- ggplot(d, aes(day, temperature)) + geom_line()
export_plot(p, "GGP-01-004")
ParameterMeaning in this example
geom_line()Connect the supplied measurements by day.

Render this diagram

tools/render examples/GGP-01-004.R

Try changing: Move the fourth reading to day 4.5.

Actual output

Connect an ordered series
GGP-01-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-01-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-01-005

Draw supplied category totals

A column layer uses the provided heights directly. Use it when counts have already been calculated rather than asking a plotting layer to count rows.

Back to diagram overview

Compare with GGP-01-004 · Connect an ordered series.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(batch=c("A","B","C"), count=c(8,14,11))
p <- ggplot(d, aes(batch, count)) + geom_col(fill="#00549F")
export_plot(p, "GGP-01-005")
ParameterMeaning in this example
geom_col()Use supplied totals as bar heights.

Render this diagram

tools/render examples/GGP-01-005.R

Try changing: Change batch B to 17.

Actual output

Draw supplied category totals
GGP-01-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-01-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-01-006

Name the question and units

Labels turn mapped columns into a chart that can stand alone. Use a clear title and units instead of leaving readers to infer the measurement.

Back to diagram overview

Compare with GGP-01-005 · Draw supplied category totals.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, temperature=c(12,14,13,16,15))
p <- ggplot(d, aes(day, temperature)) + geom_point(size=3) +
  labs(title="Archive room readings", x="Day", y="Temperature (C)")
export_plot(p, "GGP-01-006")
ParameterMeaning in this example
labs()Set the title and human-readable axis labels.

Render this diagram

tools/render examples/GGP-01-006.R

Try changing: Change the title to identify a different room.

Actual output

Name the question and units
GGP-01-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-01-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

02 · Data tables and groups

Make categories, groups and missing observations explicit before interpreting the marks.

GGP-02-001

Treat a code as a category

A factor tells the scale that values are category codes. Use it when numbers identify batches rather than measured distances.

Back to diagram overview

Compare with GGP-01-005 · Draw supplied category totals.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(batch=factor(c(1,2,3)), count=c(8,14,11))
p <- ggplot(d, aes(batch, count)) + geom_col(fill="#00549F")
export_plot(p, "GGP-02-001")
ParameterMeaning in this example
factor()Represent batch identifiers as discrete categories.

Render this diagram

tools/render examples/GGP-02-001.R

Try changing: Replace the levels with A, B and C.

Actual output

Treat a code as a category
GGP-02-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-02-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-02-002

Keep series in a long table

A long table stores one measured value per row and a group column. Use this structure when multiple stations share the same plotting rule.

Back to diagram overview

Compare with GGP-02-001 · Treat a code as a category.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=4),
                day=rep(1:4,2), reading=c(8,12,11,15,13,10,14,12))
p <- ggplot(d, aes(day, reading, colour=station)) +
  geom_point(size=3) + labs(y="Reading")
export_plot(p, "GGP-02-002")
ParameterMeaning in this example
One value columnKeep all readings in one column and identify their station separately.

Render this diagram

tools/render examples/GGP-02-002.R

Try changing: Add a third station with four observations.

Actual output

Keep series in a long table
GGP-02-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-02-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-02-003

Group connections explicitly

A group mapping says which observations belong to the same line. Use it when a continuous color or shared appearance would not establish the intended grouping.

Back to diagram overview

Compare with GGP-02-002 · Keep series in a long table.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=4),
                day=rep(1:4,2), reading=c(8,12,11,15,13,10,14,12))
p <- ggplot(d, aes(day, reading, group=station)) +
  geom_line() + geom_point(size=2)
export_plot(p, "GGP-02-003")
ParameterMeaning in this example
group=stationKeep each station's line independent.

Render this diagram

tools/render examples/GGP-02-003.R

Try changing: Remove group temporarily and inspect the connections.

Actual output

Group connections explicitly
GGP-02-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-02-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-02-004

Keep a visible gap for missing data

An NA represents an unavailable measurement rather than zero. Keep it in the sequence so the line does not imply an observed connection across the gap.

Back to diagram overview

Compare with GGP-02-003 · Group connections explicitly.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, temperature=c(12,14,NA,16,15))
p <- ggplot(d, aes(day, temperature)) +
  geom_line(na.rm=TRUE) + geom_point(na.rm=TRUE) +
  labs(caption="Day 3 was not measured")
export_plot(p, "GGP-02-004")
ParameterMeaning in this example
NAMark the unavailable reading explicitly.
na.rm=TRUEOmit missing marks while preserving the supplied gap in the line.

Render this diagram

tools/render examples/GGP-02-004.R

Try changing: Replace the missing value with 13 and update the caption.

Actual output

Keep a visible gap for missing data
GGP-02-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-02-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-02-005

Order categories deliberately

Explicit factor levels control the display order. Use a process or reporting order instead of inheriting alphabetical sorting.

Back to diagram overview

Compare with GGP-02-004 · Keep a visible gap for missing data.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(stage=c("Receive","Inspect","Archive"), count=c(20,17,14))
d$stage <- factor(d$stage, levels=c("Receive","Inspect","Archive"))
p <- ggplot(d, aes(stage, count)) + geom_col(fill="#00549F")
export_plot(p, "GGP-02-005")
ParameterMeaning in this example
levels=Set the intended left-to-right category order.

Render this diagram

tools/render examples/GGP-02-005.R

Try changing: Reverse the level order.

Actual output

Order categories deliberately
GGP-02-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-02-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-02-006

Calculate a displayed rate

A derived column makes a ratio and its denominator visible in the source. Use explicit denominators when groups have different totals.

Back to diagram overview

Compare with GGP-02-005 · Order categories deliberately.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(batch=c("A","B","C"), accepted=c(8,14,11), total=c(10,20,12))
d$percent <- 100*d$accepted/d$total
p <- ggplot(d, aes(batch, percent)) + geom_col(fill="#00549F") +
  labs(y="Accepted (%)")
export_plot(p, "GGP-02-006")
ParameterMeaning in this example
accepted / totalCalculate each batch rate with its own denominator.

Render this diagram

tools/render examples/GGP-02-006.R

Try changing: Change batch B total to 25.

Actual output

Calculate a displayed rate
GGP-02-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-02-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

03 · Layers and geometry

Add a layer only when it communicates another part of the same question.

GGP-03-001

Show samples on their trend

Points identify the observed sampling positions on a connected series. Use both layers when readers should distinguish measurements from the line between them.

Back to diagram overview

Compare with GGP-01-004 · Connect an ordered series.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, temperature=c(12,14,13,16,15))
p <- ggplot(d, aes(day, temperature)) +
  geom_line(colour="#00549F") + geom_point(size=3)
export_plot(p, "GGP-03-001")
ParameterMeaning in this example
Layer additionDraw the line first and the observed points above it.

Render this diagram

tools/render examples/GGP-03-001.R

Try changing: Increase only the point size to 4.

Actual output

Show samples on their trend
GGP-03-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-03-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-03-002

Connect paired values with segments

A segment has explicit start and end coordinates. Use it to compare two supplied readings for the same station.

Back to diagram overview

Compare with GGP-03-001 · Show samples on their trend.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=c("North","South","East"), before=c(8,14,11), after=c(12,10,16))
p <- ggplot(d, aes(y=station)) +
  geom_segment(aes(x=before,xend=after,yend=station), colour="#00549F") +
  geom_point(aes(x=before),size=3) + geom_point(aes(x=after),size=3)
export_plot(p, "GGP-03-002")
ParameterMeaning in this example
xend / yendSpecify the other endpoint of each paired comparison.

Render this diagram

tools/render examples/GGP-03-002.R

Try changing: Change the East after value to 13.

Actual output

Connect paired values with segments
GGP-03-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-03-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-03-003

Fill the area under a supplied series

An area layer relates a height to its baseline. Use it for an ordered quantity whose filled extent has a stated meaning, such as cumulative storage use.

Back to diagram overview

Compare with GGP-03-002 · Connect paired values with segments.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, used=c(2,5,8,9,12))
p <- ggplot(d, aes(day, used)) + geom_area(fill="#E8F1FA",colour="#00549F") +
  labs(y="Storage used (GB)")
export_plot(p, "GGP-03-003")
ParameterMeaning in this example
geom_area()Fill between the ordered series and zero.

Render this diagram

tools/render examples/GGP-03-003.R

Try changing: Change the last value to 15.

Actual output

Fill the area under a supplied series
GGP-03-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-03-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-03-004

Draw a stepwise state

A step layer holds a value until the next supplied change. Use it for settings or recorded state changes rather than smoothly varying measurements.

Back to diagram overview

Compare with GGP-03-003 · Fill the area under a supplied series.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=c(1,2,4,5), setting=c(10,14,12,16))
p <- ggplot(d, aes(day, setting)) + geom_step(direction="hv") +
  labs(y="Control setting")
export_plot(p, "GGP-03-004")
ParameterMeaning in this example
direction="hv"Travel horizontally before the vertical change.

Render this diagram

tools/render examples/GGP-03-004.R

Try changing: Change direction to vh and compare when each jump occurs.

Actual output

Draw a stepwise state
GGP-03-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-03-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-03-005

Show a supplied range as a band

A ribbon draws explicit lower and upper bounds. Use it when a range is known independently of a plotting layer's statistical estimator.

Back to diagram overview

Compare with GGP-03-004 · Draw a stepwise state.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, low=c(9,10,11,10,12), high=c(13,15,14,16,17))
p <- ggplot(d, aes(day)) +
  geom_ribbon(aes(ymin=low,ymax=high),fill="#E8F1FA",colour="#00549F") +
  labs(y="Supplied range")
export_plot(p, "GGP-03-005")
ParameterMeaning in this example
ymin / ymaxProvide the two boundaries of the range band.

Render this diagram

tools/render examples/GGP-03-005.R

Try changing: Increase the fourth upper bound to 18.

Actual output

Show a supplied range as a band
GGP-03-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-03-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-03-006

Close an explicit polygon

A polygon joins ordered vertices and closes the boundary. Use it for a supplied region such as a simplified field enclosure.

Back to diagram overview

Compare with GGP-03-005 · Show a supplied range as a band.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(x=c(0,3,4,2,0),y=c(0,0,2,3,2))
p <- ggplot(d, aes(x,y)) + geom_polygon(fill="#E8F1FA",colour="#00549F") +
  coord_equal() + labs(x="Field x",y="Field y")
export_plot(p, "GGP-03-006")
ParameterMeaning in this example
geom_polygon()Join the supplied vertices and close the region.

Render this diagram

tools/render examples/GGP-03-006.R

Try changing: Move the top vertex from y=3 to y=4.

Actual output

Close an explicit polygon
GGP-03-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-03-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

04 · Scales and domains

Choose mappings that match the meaning of values rather than only the available space.

GGP-04-001

Choose readable axis breaks

Explicit breaks make the reading intervals consistent. Use them when a set of related charts should share a familiar scale.

Back to diagram overview

Compare with GGP-01-001 · Place the measurements.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, temperature=c(12,14,13,16,15))
p <- ggplot(d, aes(day, temperature)) + geom_point(size=3) +
  scale_y_continuous(breaks=seq(10,18,2),limits=c(10,18))
export_plot(p, "GGP-04-001")
ParameterMeaning in this example
breaks / limitsSpecify displayed tick values and the complete scale domain.

Render this diagram

tools/render examples/GGP-04-001.R

Try changing: Change the interval to one degree.

Actual output

Choose readable axis breaks
GGP-04-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-04-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-04-002

Plot multiplicative values logarithmically

A log scale gives equal distance to equal ratios. Use it for strictly positive readings that span orders of magnitude.

Back to diagram overview

Compare with GGP-04-001 · Choose readable axis breaks.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, concentration=c(1,3,10,30,100))
p <- ggplot(d, aes(day,concentration)) + geom_point(size=3) + scale_y_log10()
export_plot(p, "GGP-04-002")
ParameterMeaning in this example
scale_y_log10()Map positive values using base-ten logarithms.

Render this diagram

tools/render examples/GGP-04-002.R

Try changing: Change the final concentration to 300.

Actual output

Plot multiplicative values logarithmically
GGP-04-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-04-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-04-003

Reverse a depth axis

A reversed scale can make increasing depth run downward. Use a unit label so the direction has a clear physical interpretation.

Back to diagram overview

Compare with GGP-04-002 · Plot multiplicative values logarithmically.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(temperature=c(15,13,11,10),depth=c(0,2,5,8))
p <- ggplot(d,aes(temperature,depth)) + geom_point(size=3) +
  scale_y_reverse() + labs(y="Depth (m)",x="Temperature (C)")
export_plot(p, "GGP-04-003")
ParameterMeaning in this example
scale_y_reverse()Place larger depth values lower in the plot.

Render this diagram

tools/render examples/GGP-04-003.R

Try changing: Add a reading at depth 10.

Actual output

Reverse a depth axis
GGP-04-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-04-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-04-004

Assign stable group colors

A named manual scale keeps a group color stable when row order changes. Use this when the same station appears in multiple figures.

Back to diagram overview

Compare with GGP-04-003 · Reverse a depth axis.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=4),
                day=rep(1:4,2), reading=c(8,12,11,15,13,10,14,12))
p <- ggplot(d,aes(day,reading,colour=station)) + geom_point(size=3) +
  scale_colour_manual(values=c(North="#00549F",South="#C45A18"))
export_plot(p, "GGP-04-004")
ParameterMeaning in this example
Named color valuesAssociate each color with its station name.

Render this diagram

tools/render examples/GGP-04-004.R

Try changing: Swap the two named colors.

Actual output

Assign stable group colors
GGP-04-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-04-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-04-005

Scale circle area to sample size

An area scale makes a larger marker represent more observations. Include zero in the scale definition so an empty count has zero area.

Back to diagram overview

Compare with GGP-04-004 · Assign stable group colors.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=c(1,3,5),reading=c(8,14,11),n=c(2,8,18))
p <- ggplot(d,aes(day,reading,size=n)) + geom_point(colour="#00549F") +
  scale_size_area(max_size=10,limits=c(0,20))
export_plot(p, "GGP-04-005")
ParameterMeaning in this example
scale_size_area()Map count to marker area with a zero-sized origin.

Render this diagram

tools/render examples/GGP-04-005.R

Try changing: Change the middle count to 12.

Actual output

Scale circle area to sample size
GGP-04-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-04-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-04-006

Map a measured value continuously

A continuous color scale represents a numeric field. Use a fixed range when the same colors must mean the same values in another chart.

Back to diagram overview

Compare with GGP-04-005 · Scale circle area to sample size.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(x=1:5,y=c(2,4,3,5,4),temperature=c(10,13,15,18,20))
p <- ggplot(d,aes(x,y,colour=temperature)) + geom_point(size=4) +
  scale_colour_gradient(low="#E8F1FA",high="#00549F",limits=c(10,20))
export_plot(p, "GGP-04-006")
ParameterMeaning in this example
limits=c(10,20)Fix the numeric meaning of the color endpoints.

Render this diagram

tools/render examples/GGP-04-006.R

Try changing: Change the fixed domain to 5 through 25.

Actual output

Map a measured value continuously
GGP-04-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-04-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

05 · Labels and themes

Use labels and presentation choices to explain the same data more clearly.

GGP-05-001

Explain the figure in its caption

A subtitle and caption can state scope without crowding the axes. Use them for the archive room and the status of the supplied observations.

Back to diagram overview

Compare with GGP-01-006 · Name the question and units.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, temperature=c(12,14,13,16,15))
p <- ggplot(d,aes(day,temperature)) + geom_point(size=3) +
  labs(title="Room temperature",subtitle="North archive",
       caption="Synthetic teaching measurements",y="Temperature (C)")
export_plot(p, "GGP-05-001")
ParameterMeaning in this example
subtitle / captionSeparate the scope from the data and provenance note.

Render this diagram

tools/render examples/GGP-05-001.R

Try changing: Change the room named in the subtitle.

Actual output

Explain the figure in its caption
GGP-05-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-05-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-05-002

Remove an unhelpful grid direction

A theme element can reduce background detail without changing the mapping. Use it when horizontal references suffice for reading the values.

Back to diagram overview

Compare with GGP-05-001 · Explain the figure in its caption.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, temperature=c(12,14,13,16,15))
p <- ggplot(d,aes(day,temperature)) + geom_point(size=3) +
  theme(panel.grid.major.x=element_blank())
export_plot(p, "GGP-05-002")
ParameterMeaning in this example
element_blank()Omit the selected theme element.

Render this diagram

tools/render examples/GGP-05-002.R

Try changing: Remove the horizontal grid instead.

Actual output

Remove an unhelpful grid direction
GGP-05-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-05-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-05-003

Rotate long category labels

A text angle makes long names fit while retaining their full wording. Use the justification to keep the rotated text attached to its tick.

Back to diagram overview

Compare with GGP-05-002 · Remove an unhelpful grid direction.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=c("North archive","South archive","East archive"),n=c(8,14,11))
p <- ggplot(d,aes(station,n)) + geom_col(fill="#00549F") +
  theme(axis.text.x=element_text(angle=35,hjust=1))
export_plot(p, "GGP-05-003")
ParameterMeaning in this example
angle / hjustRotate category text and align its end with the tick.

Render this diagram

tools/render examples/GGP-05-003.R

Try changing: Change the angle to 50 degrees.

Actual output

Rotate long category labels
GGP-05-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-05-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-05-004

Move a group legend

A legend can occupy a side instead of the common lower position. Use this when the panel needs horizontal space for a long time axis.

Back to diagram overview

Compare with GGP-05-003 · Rotate long category labels.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=4),
                day=rep(1:4,2), reading=c(8,12,11,15,13,10,14,12))
p <- ggplot(d,aes(day,reading,colour=station)) + geom_point(size=3) +
  theme(legend.position="right")
export_plot(p, "GGP-05-004")
ParameterMeaning in this example
legend.positionChoose where the mapped group legend appears.

Render this diagram

tools/render examples/GGP-05-004.R

Try changing: Move the legend to the top.

Actual output

Move a group legend
GGP-05-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-05-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-05-005

Name a legend independently

The legend title should explain the field in ordinary language. Use labs to rename it without changing the source column.

Back to diagram overview

Compare with GGP-05-004 · Move a group legend.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=4),
                day=rep(1:4,2), reading=c(8,12,11,15,13,10,14,12))
p <- ggplot(d,aes(day,reading,colour=station)) + geom_point(size=3) +
  labs(colour="Archive station")
export_plot(p, "GGP-05-005")
ParameterMeaning in this example
labs(colour=)Set the title of the color legend.

Render this diagram

tools/render examples/GGP-05-005.R

Try changing: Rename the legend to Measurement site.

Actual output

Name a legend independently
GGP-05-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-05-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-05-006

Annotate a specific position

An annotation adds explanatory text without creating a new data group. Use explicit coordinates when the note belongs to one supplied event.

Back to diagram overview

Compare with GGP-05-005 · Name a legend independently.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, temperature=c(12,14,13,16,15))
p <- ggplot(d,aes(day,temperature)) + geom_point(size=3) +
  annotate("text",x=4,y=17.5,label="Inspection",size=3.5)
export_plot(p, "GGP-05-006")
ParameterMeaning in this example
annotate()Add one explanatory label at fixed data coordinates.

Render this diagram

tools/render examples/GGP-05-006.R

Try changing: Move the note to day 2.

Actual output

Annotate a specific position
GGP-05-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-05-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

06 · Bars and rates

Distinguish counted rows, supplied totals and within-group proportions.

GGP-06-001

Count the input rows

A bar layer can count category occurrences directly. Use it when every input row is one archive record rather than a supplied total.

Back to diagram overview

Compare with GGP-01-005 · Draw supplied category totals.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(status=c("Filed","Held","Filed","Filed","Held","Filed"))
p <- ggplot(d,aes(status)) + geom_bar(fill="#00549F") + labs(y="Records")
export_plot(p, "GGP-06-001")
ParameterMeaning in this example
geom_bar()Count rows in each category.

Render this diagram

tools/render examples/GGP-06-001.R

Try changing: Add two Held records.

Actual output

Count the input rows
GGP-06-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-06-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-06-002

Rank categories by their totals

Reorder uses the supplied totals to determine category order. Use it to show a count ranking while preserving the original names.

Back to diagram overview

Compare with GGP-06-001 · Count the input rows.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(batch=c("A","B","C"),n=c(8,14,11))
p <- ggplot(d,aes(reorder(batch,n),n)) + geom_col(fill="#00549F") +
  labs(x="Batch",y="Records")
export_plot(p, "GGP-06-002")
ParameterMeaning in this example
reorder(batch,n)Arrange batch names by their supplied totals.

Render this diagram

tools/render examples/GGP-06-002.R

Try changing: Use -n to reverse the ranking.

Actual output

Rank categories by their totals
GGP-06-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-06-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-06-003

Place grouped bars side by side

A dodge position separates groups within a category. Use it when the reader needs to compare station counts for the same day.

Back to diagram overview

Compare with GGP-06-002 · Rank categories by their totals.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=rep(c("Mon","Tue"),each=2),
                station=rep(c("North","South"),2),n=c(8,12,14,10))
p <- ggplot(d,aes(day,n,fill=station)) + geom_col(position="dodge")
export_plot(p, "GGP-06-003")
ParameterMeaning in this example
position="dodge"Place the two station bars beside each other.

Render this diagram

tools/render examples/GGP-06-003.R

Try changing: Increase the South Tuesday count to 16.

Actual output

Place grouped bars side by side
GGP-06-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-06-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-06-004

Stack parts of a total

A stacked column keeps each component attached to its total. Use it when accepted and held records exhaust the supplied batch population.

Back to diagram overview

Compare with GGP-06-003 · Place grouped bars side by side.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(batch=rep(c("A","B","C"),each=2),
                status=rep(c("Accepted","Held"),3),n=c(8,2,14,3,11,1))
p <- ggplot(d,aes(batch,n,fill=status)) + geom_col() + labs(y="Records")
export_plot(p, "GGP-06-004")
ParameterMeaning in this example
fill=statusIdentify the two components of each stack.

Render this diagram

tools/render examples/GGP-06-004.R

Try changing: Change batch B Held from 3 to 5.

Actual output

Stack parts of a total
GGP-06-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-06-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-06-005

Normalize each batch to a whole

A fill position divides by each stack's total. Use it when composition is the question, and retain the original counts in the complete source.

Back to diagram overview

Compare with GGP-06-004 · Stack parts of a total.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(batch=rep(c("A","B","C"),each=2),
                status=rep(c("Accepted","Held"),3),n=c(8,2,14,3,11,1))
p <- ggplot(d,aes(batch,n,fill=status)) + geom_col(position="fill") +
  scale_y_continuous(labels=function(x) paste0(100*x,"%")) +
  labs(y="Within-batch share")
export_plot(p, "GGP-06-005")
ParameterMeaning in this example
position="fill"Normalize every batch independently to a total height of one.

Render this diagram

tools/render examples/GGP-06-005.R

Try changing: Increase a Held count and inspect the changed denominator.

Actual output

Normalize each batch to a whole
GGP-06-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-06-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-06-006

Put categories on the vertical axis

A horizontal column chart leaves room for category names. Use an explicit value mapping instead of shortening meaningful station labels.

Back to diagram overview

Compare with GGP-06-005 · Normalize each batch to a whole.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=c("North archive","South archive","East archive"),n=c(8,14,11))
p <- ggplot(d,aes(n,station)) + geom_col(fill="#00549F") + labs(x="Records",y=NULL)
export_plot(p, "GGP-06-006")
ParameterMeaning in this example
aes(n,station)Place the numeric value horizontally and category vertically.

Render this diagram

tools/render examples/GGP-06-006.R

Try changing: Order the stations by n.

Actual output

Put categories on the vertical axis
GGP-06-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-06-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

07 · Distributions

State how bins, density estimates and summaries are constructed from the observations.

GGP-07-001

Fix histogram bins explicitly

A histogram counts observations in intervals. Use a stated width, boundary and closure so the same values cannot silently move between bins.

Back to diagram overview

Compare with GGP-06-001 · Count the input rows.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(value=c(8,9,9,10,11,12,12,13,14,15,16,18))
p <- ggplot(d,aes(value)) +
  geom_histogram(binwidth=2,boundary=0,closed="left",
                 fill="#00549F",colour="white")
export_plot(p, "GGP-07-001")
ParameterMeaning in this example
binwidth / boundary / closedUse two-unit intervals aligned to zero and closed on the left.

Render this diagram

tools/render examples/GGP-07-001.R

Try changing: Change the bin width to 3.

Actual output

Fix histogram bins explicitly
GGP-07-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-07-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-07-002

Draw counts as a frequency polygon

A frequency polygon joins bin-center counts. Use the same bin definition as a histogram when comparing the two geometries.

Back to diagram overview

Compare with GGP-07-001 · Fix histogram bins explicitly.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(value=c(8,9,9,10,11,12,12,13,14,15,16,18))
p <- ggplot(d,aes(value)) +
  geom_freqpoly(binwidth=2,boundary=0,closed="left",colour="#00549F")
export_plot(p, "GGP-07-002")
ParameterMeaning in this example
geom_freqpoly()Join the supplied binning rule's center counts.

Render this diagram

tools/render examples/GGP-07-002.R

Try changing: Change the width to 3 and compare the shape.

Actual output

Draw counts as a frequency polygon
GGP-07-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-07-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-07-003

Fix a density bandwidth

A density layer estimates a smooth distribution rather than counts. Use an explicit bandwidth so the smoothing choice is visible.

Back to diagram overview

Compare with GGP-07-002 · Draw counts as a frequency polygon.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(value=c(8,9,9,10,11,12,12,13,14,15,16,18))
p <- ggplot(d,aes(value)) +
  geom_density(bw=1,fill="#E8F1FA",colour="#00549F")
export_plot(p, "GGP-07-003")
ParameterMeaning in this example
bw=1Use a Gaussian smoothing bandwidth of one data unit.

Render this diagram

tools/render examples/GGP-07-003.R

Try changing: Compare bandwidths .5 and 2.

Actual output

Fix a density bandwidth
GGP-07-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-07-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-07-004

Show the cumulative distribution

An empirical cumulative curve reports the fraction at or below each value. Use it to compare quantiles without selecting histogram bins.

Back to diagram overview

Compare with GGP-07-003 · Fix a density bandwidth.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(value=c(8,9,9,10,11,12,12,13,14,15,16,18))
p <- ggplot(d,aes(value)) + stat_ecdf(geom="step",pad=FALSE) +
  labs(y="Cumulative fraction")
export_plot(p, "GGP-07-004")
ParameterMeaning in this example
stat_ecdf()Compute the empirical cumulative distribution including ties.

Render this diagram

tools/render examples/GGP-07-004.R

Try changing: Add another observation at value 12.

Actual output

Show the cumulative distribution
GGP-07-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-07-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-07-005

Expose the boxplot whisker rule

A native boxplot estimates quartiles and flags distant observations. Here whiskers extend to observations within 1.5 interquartile ranges; they are not automatically the minimum and maximum.

Back to diagram overview

Compare with GGP-07-004 · Show the cumulative distribution.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=8),
                value=c(9,10,11,11,12,13,14,24,10,11,12,13,14,15,16,17))
p <- ggplot(d,aes(station,value)) +
  geom_boxplot(coef=1.5,fill="#E8F1FA",colour="#00549F")
export_plot(p, "GGP-07-005")
ParameterMeaning in this example
coef=1.5Set the interquartile-range multiplier for whisker limits.

Render this diagram

tools/render examples/GGP-07-005.R

Try changing: Change coef to 2 and inspect the North outlier.

Actual output

Expose the boxplot whisker rule
GGP-07-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-07-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-07-006

Compare native violin densities

A violin mirrors a density estimate within each category. Use a common bandwidth and equal-area scaling when comparing station distributions.

Back to diagram overview

Compare with GGP-07-005 · Expose the boxplot whisker rule.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=8),
                value=c(9,10,11,11,12,13,14,24,10,11,12,13,14,15,16,17))
p <- ggplot(d,aes(station,value)) +
  geom_violin(bw=1,trim=FALSE,scale="area",fill="#E8F1FA")
export_plot(p, "GGP-07-006")
ParameterMeaning in this example
trim=FALSE / scale="area"Retain estimated tails and normalize the violin areas.

Render this diagram

tools/render examples/GGP-07-006.R

Try changing: Change the bandwidth to 2.

Actual output

Compare native violin densities
GGP-07-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-07-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

08 · Uncertainty and summaries

Show where every center and interval comes from; a range needs a stated definition.

GGP-08-001

Draw supplied measurement bounds

An error bar can display bounds provided with a reading. Use explicit lower and upper columns rather than implying an unstated confidence procedure.

Back to diagram overview

Compare with GGP-03-005 · Show a supplied range as a band.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=c("North","South","East"),mean=c(11,14,12),
                low=c(9,12,10),high=c(13,17,15))
p <- ggplot(d,aes(station,mean)) + geom_point(size=3) +
  geom_errorbar(aes(ymin=low,ymax=high),width=.2) +
  labs(y="Reading with supplied bounds")
export_plot(p, "GGP-08-001")
ParameterMeaning in this example
ymin / ymaxRead the supplied interval columns directly.

Render this diagram

tools/render examples/GGP-08-001.R

Try changing: Widen the East upper bound to 17.

Actual output

Draw supplied measurement bounds
GGP-08-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-08-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-08-002

Summarize raw groups with a layer

A summary layer can calculate a center from the supplied observations. Here the interval is the observed range, with no claim of statistical confidence.

Back to diagram overview

Compare with GGP-08-001 · Draw supplied measurement bounds.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=8),
                value=c(9,10,11,11,12,13,14,24,10,11,12,13,14,15,16,17))
p <- ggplot(d,aes(station,value)) +
  stat_summary(fun=mean,fun.min=min,fun.max=max,geom="pointrange") +
  labs(y="Mean and observed range")
export_plot(p, "GGP-08-002")
ParameterMeaning in this example
fun / fun.min / fun.maxName the three explicit summary functions.

Render this diagram

tools/render examples/GGP-08-002.R

Try changing: Replace mean with median.

Actual output

Summarize raw groups with a layer
GGP-08-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-08-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-08-003

Summarize observations in fixed bins

A binned summary groups a continuous coordinate into intervals. Use a stated bin width when irregularly spaced observations need a comparable local mean.

Back to diagram overview

Compare with GGP-08-002 · Summarize raw groups with a layer.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=c(.2,.7,1.1,2.4,2.8,3.3,4.1,4.8,5.2),
                reading=c(8,10,9,13,15,12,14,16,15))
p <- ggplot(d,aes(day,reading)) + geom_point(colour="grey60") +
  stat_summary_bin(fun=mean,geom="point",binwidth=2,colour="#00549F",size=4)
export_plot(p, "GGP-08-003")
ParameterMeaning in this example
stat_summary_bin()Calculate the supplied function within each fixed-width bin.

Render this diagram

tools/render examples/GGP-08-003.R

Try changing: Change binwidth to 1.

Actual output

Summarize observations in fixed bins
GGP-08-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-08-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-08-004

State a weighted mean explicitly

Weights change the contribution of each supplied value. Use an explicit calculation and caption when weights represent known record multiplicities.

Back to diagram overview

Compare with GGP-08-003 · Summarize observations in fixed bins.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(batch=c("A","B","C"),reading=c(8,14,11),weight=c(1,3,2))
m <- weighted.mean(d$reading,d$weight)
p <- ggplot(d,aes(batch,reading)) + geom_point(size=3) +
  geom_hline(yintercept=m,linetype="dashed") +
  labs(caption="Dashed line: mean weighted by record multiplicity")
export_plot(p, "GGP-08-004")
ParameterMeaning in this example
weighted.mean()Calculate the center using the supplied multiplicities.

Render this diagram

tools/render examples/GGP-08-004.R

Try changing: Change batch B weight to 1.

Actual output

State a weighted mean explicitly
GGP-08-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-08-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-08-005

Calculate one standard error

A standard error describes variability of the sample mean under the stated sampling interpretation. This example displays one standard error on each side of the mean, not a confidence interval.

Back to diagram overview

Compare with GGP-08-004 · State a weighted mean explicitly.

Complete R / ggplot2 source

source("config/style.R")
values <- c(9,10,11,11,12,13,14,16)
m <- mean(values); se <- sd(values)/sqrt(length(values))
d <- data.frame(group="North",mean=m,low=m-se,high=m+se)
p <- ggplot(d,aes(group,mean,ymin=low,ymax=high)) +
  geom_pointrange(colour="#00549F") + labs(y="Mean +/- one standard error")
export_plot(p, "GGP-08-005")
ParameterMeaning in this example
sd(x)/sqrt(n)Calculate the standard error of the supplied sample mean.

Render this diagram

tools/render examples/GGP-08-005.R

Try changing: Add another value of 16 and recalculate.

Actual output

Calculate one standard error
GGP-08-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-08-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-08-006

Make a reproducible bootstrap interval

Resampling can provide an illustrative percentile interval for the mean. Use an explicit seed, number of resamples and quantile rule so the computational procedure can be reproduced.

Back to diagram overview

Compare with GGP-08-005 · Calculate one standard error.

Complete R / ggplot2 source

source("config/style.R")
RNGkind("Mersenne-Twister","Inversion","Rejection"); set.seed(17)
values <- c(9,10,11,11,12,13,14,16)
means <- replicate(400,mean(sample(values,replace=TRUE)))
bounds <- quantile(means,c(.025,.975),type=7)
d <- data.frame(group="North",mean=mean(values),low=bounds[1],high=bounds[2])
p <- ggplot(d,aes(group,mean,ymin=low,ymax=high)) +
  geom_pointrange(colour="#00549F") +
  labs(y="Mean and percentile bootstrap interval",caption="400 resamples; seed 17")
export_plot(p, "GGP-08-006")
ParameterMeaning in this example
replicate / sampleGenerate the stated number of replacement samples.
type=7Use an explicit interpolated quantile rule.

Render this diagram

tools/render examples/GGP-08-006.R

Try changing: Increase the resample count to 800 and update the caption.

Actual output

Make a reproducible bootstrap interval
GGP-08-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-08-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

09 · Relationships and models

Separate the supplied observations from the model or aggregation drawn over them.

GGP-09-001

Fit a stated linear trend

A smoothing layer can fit an explicit least-squares model. Keep the observations visible and state the model instead of implying that the line is another measured series.

Back to diagram overview

Compare with GGP-03-001 · Show samples on their trend.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(x=1:7,y=c(8,10,9,13,14,13,17))
p <- ggplot(d,aes(x,y)) + geom_point(size=3) +
  geom_smooth(method="lm",formula=y~x,se=FALSE,colour="#00549F")
export_plot(p, "GGP-09-001")
ParameterMeaning in this example
method="lm" / formula=y~xFit the stated linear mean model.

Render this diagram

tools/render examples/GGP-09-001.R

Try changing: Change one observation and inspect the fitted line.

Actual output

Fit a stated linear trend
GGP-09-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-09-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-09-002

Show uncertainty in the fitted mean

The smooth layer can draw a confidence band for the model mean. This is a model-based interval under linear-model assumptions, not a prediction interval for individual readings.

Back to diagram overview

Compare with GGP-09-001 · Fit a stated linear trend.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(x=1:7,y=c(8,10,9,13,14,13,17))
p <- ggplot(d,aes(x,y)) + geom_point(size=3) +
  geom_smooth(method="lm",formula=y~x,se=TRUE,level=.95)
export_plot(p, "GGP-09-002")
ParameterMeaning in this example
se / levelDisplay the stated confidence level for the fitted mean.

Render this diagram

tools/render examples/GGP-09-002.R

Try changing: Change level to .8.

Actual output

Show uncertainty in the fitted mean
GGP-09-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-09-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-09-003

Fit a quadratic reference curve

An explicit polynomial formula changes the model while leaving the observations intact. Use it for an illustrative curved relationship with enough data to inspect the fit.

Back to diagram overview

Compare with GGP-09-002 · Show uncertainty in the fitted mean.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(x=0:6,y=c(3,5,8,14,20,30,40))
p <- ggplot(d,aes(x,y)) + geom_point(size=3) +
  geom_smooth(method="lm",formula=y~poly(x,2,raw=TRUE),se=FALSE)
export_plot(p, "GGP-09-003")
ParameterMeaning in this example
poly(x,2,raw=TRUE)Fit an intercept, linear term and squared term.

Render this diagram

tools/render examples/GGP-09-003.R

Try changing: Change the middle observation from 14 to 20.

Actual output

Fit a quadratic reference curve
GGP-09-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-09-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-09-004

Report the correlation explicitly

A computed correlation can accompany a scatterplot without replacing it. Use a caption that avoids interpreting this descriptive association as a causal effect.

Back to diagram overview

Compare with GGP-09-003 · Fit a quadratic reference curve.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(x=1:7,y=c(8,10,9,13,14,13,17))
r <- cor(d$x,d$y)
p <- ggplot(d,aes(x,y)) + geom_point(size=3) +
  labs(subtitle=sprintf("Pearson r = %.2f",r),
       caption="Descriptive association in synthetic observations")
export_plot(p, "GGP-09-004")
ParameterMeaning in this example
cor()Calculate the displayed Pearson correlation from the same rows.

Render this diagram

tools/render examples/GGP-09-004.R

Try changing: Change the seventh y value to 8.

Actual output

Report the correlation explicitly
GGP-09-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-09-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-09-005

Count observations in two-dimensional bins

A binned layer aggregates overlapping observations into cells. Use explicit bin widths and a count legend so color has a defined meaning.

Back to diagram overview

Compare with GGP-09-004 · Report the correlation explicitly.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(x=c(1,1.2,1.5,2.4,2.7,3.1,3.2,4.2,4.5,5.1,5.4,5.6),
                y=c(2,2.2,2.5,3.8,4.1,3.5,4.2,5.1,5.3,6.1,5.8,6.4))
p <- ggplot(d,aes(x,y)) + geom_bin_2d(binwidth=c(2,2)) +
  labs(fill="Observations")
export_plot(p, "GGP-09-005")
ParameterMeaning in this example
binwidth=c(2,2)Count rows within the specified rectangular data cells.

Render this diagram

tools/render examples/GGP-09-005.R

Try changing: Change both bin widths to one.

Actual output

Count observations in two-dimensional bins
GGP-09-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-09-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-09-006

Expose coincident observations

A count layer combines observations at identical coordinates. Use it when repeated discrete readings would otherwise hide under one point.

Back to diagram overview

Compare with GGP-09-005 · Count observations in two-dimensional bins.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(x=c(1,1,1,2,2,3,3,3,3),y=c(2,2,2,3,3,4,4,4,4))
p <- ggplot(d,aes(x,y)) + geom_count(colour="#00549F") +
  labs(size="Repeated observations")
export_plot(p, "GGP-09-006")
ParameterMeaning in this example
geom_count()Represent coordinate multiplicity with point area.

Render this diagram

tools/render examples/GGP-09-006.R

Try changing: Add another row at 2,3.

Actual output

Expose coincident observations
GGP-09-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-09-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

10 · Facets and panels

Use panels to separate groups while making the scale-sharing decision explicit.

GGP-10-001

Give each station its own panel

A wrapped facet splits a long table by station. Use common scales when the heights should remain directly comparable.

Back to diagram overview

Compare with GGP-02-002 · Keep series in a long table.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=4),
                day=rep(1:4,2), reading=c(8,12,11,15,13,10,14,12))
p <- ggplot(d,aes(day,reading)) + geom_line() + geom_point() +
  facet_wrap(vars(station))
export_plot(p, "GGP-10-001")
ParameterMeaning in this example
facet_wrap()Arrange one panel for each station.

Render this diagram

tools/render examples/GGP-10-001.R

Try changing: Change the wrap to one column.

Actual output

Give each station its own panel
GGP-10-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-10-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-10-002

Facet by two grouping variables

A facet grid assigns one variable to rows and another to columns. Use it to make both station and measurement phase visible.

Back to diagram overview

Compare with GGP-10-001 · Give each station its own panel.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=6),
                phase=rep(rep(c("Before","After"),each=3),2),
                day=rep(1:3,4),reading=c(8,10,9,12,14,13,11,13,12,15,16,14))
p <- ggplot(d,aes(day,reading)) + geom_point(size=2) +
  facet_grid(station~phase)
export_plot(p, "GGP-10-002")
ParameterMeaning in this example
facet_grid(station~phase)Cross station rows with phase columns.

Render this diagram

tools/render examples/GGP-10-002.R

Try changing: Exchange the row and column variables.

Actual output

Facet by two grouping variables
GGP-10-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-10-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-10-003

Allow separate value scales

Free vertical scales emphasize within-panel variation. State this choice because equal heights in the two panels no longer mean equal values.

Back to diagram overview

Compare with GGP-10-002 · Facet by two grouping variables.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=4),
                day=rep(1:4,2), reading=c(8,12,11,15,13,10,14,12))
d$reading[d$station=="South"] <- 10*d$reading[d$station=="South"]
p <- ggplot(d,aes(day,reading)) + geom_line() +
  facet_wrap(vars(station),scales="free_y")
export_plot(p, "GGP-10-003")
ParameterMeaning in this example
scales="free_y"Fit a separate vertical range in each panel.

Render this diagram

tools/render examples/GGP-10-003.R

Try changing: Return scales to fixed.

Actual output

Allow separate value scales
GGP-10-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-10-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-10-004

Write readable panel labels

A labeller can expand stored group codes for readers. Use it to preserve concise data identifiers while explaining the station in the figure.

Back to diagram overview

Compare with GGP-10-003 · Allow separate value scales.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=4),
                day=rep(1:4,2), reading=c(8,12,11,15,13,10,14,12))
p <- ggplot(d,aes(day,reading)) + geom_point(size=2) +
  facet_wrap(vars(station),labeller=labeller(station=c(
    North="North archive room",South="South archive room")))
export_plot(p, "GGP-10-004")
ParameterMeaning in this example
labeller()Map stored station values to readable strip labels.

Render this diagram

tools/render examples/GGP-10-004.R

Try changing: Change the two display labels.

Actual output

Write readable panel labels
GGP-10-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-10-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-10-005

Retain an unmeasured category

An unused factor level can remain as an explicitly empty panel. Use it to show that a planned station has no observations, rather than implying that its reading was zero.

Back to diagram overview

Compare with GGP-10-004 · Write readable panel labels.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=4),
                day=rep(1:4,2), reading=c(8,12,11,15,13,10,14,12))
d$station <- factor(d$station,levels=c("North","South","East"))
p <- ggplot(d,aes(day,reading)) + geom_point(size=2) +
  facet_wrap(vars(station),drop=FALSE) +
  labs(caption="East was not measured in this fixture")
export_plot(p, "GGP-10-005")
ParameterMeaning in this example
drop=FALSERetain the declared category with no observed rows.

Render this diagram

tools/render examples/GGP-10-005.R

Try changing: Set drop to TRUE and compare the inventory.

Actual output

Retain an unmeasured category
GGP-10-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-10-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-10-006

Repeat one reference across panels

A layer with a fixed reference can appear in every facet. Use it when all stations share the same operational target.

Back to diagram overview

Compare with GGP-10-005 · Retain an unmeasured category.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=4),
                day=rep(1:4,2), reading=c(8,12,11,15,13,10,14,12))
p <- ggplot(d,aes(day,reading)) + geom_line() +
  geom_hline(yintercept=12,linetype="dashed") +
  facet_wrap(vars(station)) + labs(caption="Dashed target: 12")
export_plot(p, "GGP-10-006")
ParameterMeaning in this example
Fixed reference layerRepeat the stated target in each station panel.

Render this diagram

tools/render examples/GGP-10-006.R

Try changing: Change the target to 14 and update the caption.

Actual output

Repeat one reference across panels
GGP-10-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-10-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

11 · Coordinates and reference geometry

Distinguish a viewing window from a data transformation or analytical reference.

GGP-11-001

Zoom without discarding fit inputs

A coordinate window changes what is visible after the model is fitted. Use it when the chart needs a closer view but the analysis should still use every supplied observation.

Back to diagram overview

Compare with GGP-04-001 · Choose readable axis breaks.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, temperature=c(12,14,13,16,15))
p <- ggplot(d,aes(day,temperature)) + geom_point(size=3) +
  geom_smooth(method="lm",formula=y~x,se=FALSE) +
  coord_cartesian(xlim=c(2,4),ylim=c(12,17))
export_plot(p, "GGP-11-001")
ParameterMeaning in this example
coord_cartesian()Zoom after the layers and model have been calculated.

Render this diagram

tools/render examples/GGP-11-001.R

Try changing: Widen the horizontal window to 1 through 5.

Actual output

Zoom without discarding fit inputs
GGP-11-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-11-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-11-002

Preserve equal spatial units

A fixed coordinate ratio makes one unit equally long on both axes. Use it for a field layout where geometric shape has a physical meaning.

Back to diagram overview

Compare with GGP-11-001 · Zoom without discarding fit inputs.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(x=c(0,3,4,2,0),y=c(0,0,2,3,2))
p <- ggplot(d,aes(x,y)) + geom_polygon(fill="#E8F1FA",colour="#00549F") +
  coord_fixed(ratio=1) + labs(x="East (m)",y="North (m)")
export_plot(p, "GGP-11-002")
ParameterMeaning in this example
ratio=1Use equal physical lengths for one horizontal and vertical data unit.

Render this diagram

tools/render examples/GGP-11-002.R

Try changing: Change the ratio to .5 and inspect the distortion.

Actual output

Preserve equal spatial units
GGP-11-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-11-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-11-003

Lay out categories around a circle

Polar coordinates transform the completed rectangular layers. Use them for a cyclic display only when readers can still interpret the category order and radial values.

Back to diagram overview

Compare with GGP-11-002 · Preserve equal spatial units.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(period=factor(c("Morning","Noon","Evening","Night"),
                levels=c("Morning","Noon","Evening","Night")),n=c(8,14,11,5))
p <- ggplot(d,aes(period,n)) + geom_col(fill="#00549F") + coord_polar()
export_plot(p, "GGP-11-003")
ParameterMeaning in this example
coord_polar()Place category positions around the circle and counts radially.

Render this diagram

tools/render examples/GGP-11-003.R

Try changing: Change the Night count to 9.

Actual output

Lay out categories around a circle
GGP-11-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-11-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-11-004

Transform coordinates after calculation

A coordinate transformation changes the drawn scale after layer calculations. Use an explicit transform when that order differs from transforming the input data.

Back to diagram overview

Compare with GGP-11-003 · Lay out categories around a circle.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5,count=c(1,4,9,16,25))
p <- ggplot(d,aes(day,count)) + geom_point(size=3) +
  coord_transform(y="sqrt")
export_plot(p, "GGP-11-004")
ParameterMeaning in this example
coord_transform(y="sqrt")Apply the square-root coordinate transformation to the completed layer.

Render this diagram

tools/render examples/GGP-11-004.R

Try changing: Change the transform to log10, keeping all counts positive.

Actual output

Transform coordinates after calculation
GGP-11-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-11-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-11-005

Put a note outside the panel

Turning clipping off allows an annotation beyond a fixed viewing window. Add a margin so the exported page still contains the complete note.

Back to diagram overview

Compare with GGP-11-004 · Transform coordinates after calculation.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, temperature=c(12,14,13,16,15))
p <- ggplot(d,aes(day,temperature)) + geom_point(size=3) +
  annotate("text",x=5.3,y=15,label="End",hjust=0) +
  coord_cartesian(xlim=c(1,5),clip="off") +
  theme(plot.margin=margin(12,70,12,12))
export_plot(p, "GGP-11-005")
ParameterMeaning in this example
clip="off"Allow the annotation outside the panel window.
plot.marginReserve real page space for the external note.

Render this diagram

tools/render examples/GGP-11-005.R

Try changing: Change End to Final and inspect the page edge.

Actual output

Put a note outside the panel
GGP-11-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-11-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-11-006

Compare calibration with an identity line

An identity line marks equal supplied and measured values. Use equal axes so distance from that reference is not distorted by different physical scales.

Back to diagram overview

Compare with GGP-11-005 · Put a note outside the panel.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(reference=c(5,10,15,20),observed=c(6,9,16,19))
p <- ggplot(d,aes(reference,observed)) + geom_point(size=3) +
  geom_abline(slope=1,intercept=0,linetype="dashed") +
  coord_equal(xlim=c(0,22),ylim=c(0,22))
export_plot(p, "GGP-11-006")
ParameterMeaning in this example
slope=1 / intercept=0Draw the equality reference, not a fitted regression.

Render this diagram

tools/render examples/GGP-11-006.R

Try changing: Change one observed value to match its reference.

Actual output

Compare calibration with an identity line
GGP-11-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-11-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

12 · Dates and time series

Use real date/time classes and fixed time zones instead of formatting arbitrary numeric positions.

GGP-12-001

Use a calendar date axis

A Date column provides calendar positions with native date labels. Use it when elapsed days must retain their actual dates.

Back to diagram overview

Compare with GGP-01-004 · Connect an ordered series.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=as.Date("2025-03-01")+0:4,temperature=c(12,14,13,16,15))
p <- ggplot(d,aes(day,temperature)) + geom_line() + geom_point() +
  labs(x="Date",y="Temperature (C)")
export_plot(p, "GGP-12-001")
ParameterMeaning in this example
as.Date()Create fixed calendar dates rather than numeric category codes.

Render this diagram

tools/render examples/GGP-12-001.R

Try changing: Move the starting date to April 1.

Actual output

Use a calendar date axis
GGP-12-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-12-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-12-002

Choose date tick intervals

A date scale can set both interval and displayed format. Use a fixed interval to keep labels legible across a short observation period.

Back to diagram overview

Compare with GGP-12-001 · Use a calendar date axis.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=as.Date("2025-03-01")+0:4,temperature=c(12,14,13,16,15))
p <- ggplot(d,aes(day,temperature)) + geom_line() +
  scale_x_date(date_breaks="2 days",date_labels="%d %b")
export_plot(p, "GGP-12-002")
ParameterMeaning in this example
date_breaks / date_labelsSet the calendar interval and readable date format.

Render this diagram

tools/render examples/GGP-12-002.R

Try changing: Change the tick interval to one day.

Actual output

Choose date tick intervals
GGP-12-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-12-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-12-003

Pin timestamps to UTC

A timestamp includes a time zone as well as a date. Use an explicit zone for construction and display so the same source does not shift between machines.

Back to diagram overview

Compare with GGP-12-002 · Choose date tick intervals.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(time=as.POSIXct("2025-03-01 08:00",tz="UTC")+(0:4)*3600,
                reading=c(12,14,13,16,15))
p <- ggplot(d,aes(time,reading)) + geom_line() +
  scale_x_datetime(date_breaks="2 hours",date_labels="%H:%M",timezone="UTC") +
  labs(x="Time (UTC)")
export_plot(p, "GGP-12-003")
ParameterMeaning in this example
tz / timezoneUse UTC for both stored timestamps and tick labels.

Render this diagram

tools/render examples/GGP-12-003.R

Try changing: Change the starting time to 06:00.

Actual output

Pin timestamps to UTC
GGP-12-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-12-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-12-004

Complete a timeline before plotting

An absent row is different from an explicit missing value. Merge the observations onto a complete daily calendar so the line exposes the unmeasured day.

Back to diagram overview

Compare with GGP-12-003 · Pin timestamps to UTC.

Complete R / ggplot2 source

source("config/style.R")
observed <- data.frame(day=as.Date("2025-03-01")+c(0,1,3,4),value=c(12,14,16,15))
calendar <- data.frame(day=seq(min(observed$day),max(observed$day),by="day"))
d <- merge(calendar,observed,by="day",all.x=TRUE)
p <- ggplot(d,aes(day,value)) + geom_line(na.rm=TRUE) +
  geom_point(na.rm=TRUE) + labs(caption="March 3 was not measured")
export_plot(p, "GGP-12-004")
ParameterMeaning in this example
all.x=TRUERetain every calendar date even when its value is missing.

Render this diagram

tools/render examples/GGP-12-004.R

Try changing: Add an observation for March 3.

Actual output

Complete a timeline before plotting
GGP-12-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-12-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-12-005

Mark a dated event

A vertical reference can identify a known event on the calendar axis. Use a caption to name the event instead of implying it explains all nearby changes.

Back to diagram overview

Compare with GGP-12-004 · Complete a timeline before plotting.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=as.Date("2025-03-01")+0:4,temperature=c(12,14,13,16,15))
p <- ggplot(d,aes(day,temperature)) + geom_line() +
  geom_vline(xintercept=as.Date("2025-03-03"),linetype="dashed") +
  labs(caption="Dashed line: scheduled inspection")
export_plot(p, "GGP-12-005")
ParameterMeaning in this example
Date-valued xinterceptPlace the event on the same calendar scale as the observations.

Render this diagram

tools/render examples/GGP-12-005.R

Try changing: Move the inspection to March 4.

Actual output

Mark a dated event
GGP-12-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-12-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-12-006

Align station calendars

Facets can share a real date axis across groups. Use fixed date scales when the timing of observations should be compared directly.

Back to diagram overview

Compare with GGP-12-005 · Mark a dated event.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=4),
                day=rep(1:4,2), reading=c(8,12,11,15,13,10,14,12))
d$date <- as.Date("2025-03-01")+d$day-1
p <- ggplot(d,aes(date,reading)) + geom_line() + geom_point() +
  facet_grid(station~.) + scale_x_date(date_labels="%d %b")
export_plot(p, "GGP-12-006")
ParameterMeaning in this example
Shared date facetsCompare station timelines using the same calendar mapping.

Render this diagram

tools/render examples/GGP-12-006.R

Try changing: Move the South dates forward by one day.

Actual output

Align station calendars
GGP-12-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-12-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

13 · Matrices and fields

State whether a cell is measured, estimated or absent, and preserve the meaning of its color scale.

GGP-13-001

Draw a measured cell grid

A tile layer gives each supplied cell a boundary. Use explicit coordinates when the grid represents field positions rather than an unordered table.

Back to diagram overview

Compare with GGP-04-006 · Map a measured value continuously.

Complete R / ggplot2 source

source("config/style.R")
d <- expand.grid(x=1:5,y=1:4)
d$value <- 3*d$x+2*d$y
p <- ggplot(d,aes(x,y,fill=value)) + geom_tile(width=.95,height=.95) +
  coord_equal() + scale_fill_gradient(low="#E8F1FA",high="#00549F")
export_plot(p, "GGP-13-001")
ParameterMeaning in this example
geom_tile()Draw one rectangular cell per supplied coordinate.

Render this diagram

tools/render examples/GGP-13-001.R

Try changing: Change the gap by setting both tile dimensions to .85.

Actual output

Draw a measured cell grid
GGP-13-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-13-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-13-002

Render a regular field as a raster

A raster layer is efficient for a complete regular grid. Its cell image is an intentional raster layer inside the native PDF and SVG, with vector axes retained.

Back to diagram overview

Compare with GGP-13-001 · Draw a measured cell grid.

Complete R / ggplot2 source

source("config/style.R")
d <- expand.grid(x=seq(-2,2,length.out=31),y=seq(-2,2,length.out=31))
d$value <- exp(-d$x^2-d$y^2)
p <- ggplot(d,aes(x,y,fill=value)) + geom_raster(interpolate=FALSE) +
  coord_equal() + scale_fill_gradient(low="#E8F1FA",high="#00549F")
export_plot(p, "GGP-13-002")
ParameterMeaning in this example
interpolate=FALSEDisplay the supplied regular cells without visual interpolation.

Render this diagram

tools/render examples/GGP-13-002.R

Try changing: Use 51 grid positions in each direction.

Actual output

Render a regular field as a raster
GGP-13-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-13-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-13-003

Distinguish a missing cell from zero

A missing value needs its own appearance. Use a neutral color and a caption so an unmeasured cell is not read as the scale minimum.

Back to diagram overview

Compare with GGP-13-002 · Render a regular field as a raster.

Complete R / ggplot2 source

source("config/style.R")
d <- expand.grid(x=1:5,y=1:4)
d$value <- 3*d$x+2*d$y
d$value[d$x==3 & d$y==2] <- NA
p <- ggplot(d,aes(x,y,fill=value)) + geom_tile() + coord_equal() +
  scale_fill_gradient(low="#E8F1FA",high="#00549F",na.value="#ECECED") +
  labs(caption="Grey cell: not measured")
export_plot(p, "GGP-13-003")
ParameterMeaning in this example
na.valueGive an unavailable observation a separate fill.

Render this diagram

tools/render examples/GGP-13-003.R

Try changing: Move the missing cell to x=2,y=3.

Actual output

Distinguish a missing cell from zero
GGP-13-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-13-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-13-004

Trace explicit contour levels

Contours join positions where a supplied field has the same value. Use explicit levels so the boundaries have a reproducible meaning.

Back to diagram overview

Compare with GGP-13-003 · Distinguish a missing cell from zero.

Complete R / ggplot2 source

source("config/style.R")
d <- expand.grid(x=seq(-2,2,length.out=41),y=seq(-2,2,length.out=41))
d$value <- d$x^2+d$y^2
p <- ggplot(d,aes(x,y,z=value)) +
  geom_contour(breaks=c(.5,1,2,3),colour="#00549F") + coord_equal()
export_plot(p, "GGP-13-004")
ParameterMeaning in this example
z / breaksProvide the field values and the exact contour levels.

Render this diagram

tools/render examples/GGP-13-004.R

Try changing: Add a contour at value 1.5.

Actual output

Trace explicit contour levels
GGP-13-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-13-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-13-005

Fill a field between contour levels

Filled contours classify a continuous field into stated bands. Use a sequential palette so the bands retain their order.

Back to diagram overview

Compare with GGP-13-004 · Trace explicit contour levels.

Complete R / ggplot2 source

source("config/style.R")
d <- expand.grid(x=seq(-2,2,length.out=31),y=seq(-2,2,length.out=31))
d$value <- exp(-d$x^2-d$y^2)
p <- ggplot(d,aes(x,y,z=value)) +
  geom_contour_filled(breaks=seq(0,1,.2)) + coord_equal() +
  scale_fill_brewer(palette="Blues") + labs(fill="Field band")
export_plot(p, "GGP-13-005")
ParameterMeaning in this example
geom_contour_filled()Fill the regions between the supplied field levels.

Render this diagram

tools/render examples/GGP-13-005.R

Try changing: Change the band interval to .25.

Actual output

Fill a field between contour levels
GGP-13-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-13-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-13-006

Keep field polygons separate

A group mapping closes each supplied region independently. Use it for schematic field sections without implying that these coordinates are a geographic map.

Back to diagram overview

Compare with GGP-13-005 · Fill a field between contour levels.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(section=rep(c("A","B"),each=4),
                x=c(0,2,2,0,2,5,5,2),y=c(0,0,3,3,0,0,3,3))
p <- ggplot(d,aes(x,y,group=section,fill=section)) +
  geom_polygon(colour="#00549F") + coord_equal()
export_plot(p, "GGP-13-006")
ParameterMeaning in this example
group=sectionClose each region without connecting it to another region.

Render this diagram

tools/render examples/GGP-13-006.R

Try changing: Move the outer edge of section B from x=5 to x=6.

Actual output

Keep field polygons separate
GGP-13-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-13-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

14 · Annotation and composition

Attach explanatory marks to the data and reserve enough room for them in the finished figure.

GGP-14-001

Label the observations directly

A text layer can read identifiers from the same rows as the points. Use a small position adjustment and enough scale expansion to keep every label readable.

Back to diagram overview

Compare with GGP-05-006 · Annotate a specific position.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=c(1,3,5),reading=c(8,14,11),id=c("R17","R18","R19"))
p <- ggplot(d,aes(day,reading)) + geom_point(size=3) +
  geom_text(aes(label=id),nudge_y=.5) +
  scale_y_continuous(expand=expansion(mult=c(.1,.2)))
export_plot(p, "GGP-14-001")
ParameterMeaning in this example
aes(label=id)Read the direct labels from the observation table.

Render this diagram

tools/render examples/GGP-14-001.R

Try changing: Change the three record identifiers.

Actual output

Label the observations directly
GGP-14-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-14-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-14-002

Name each series at its end

A layer can use only the last observation of each series. Use direct endpoint labels when they make a separate group legend unnecessary.

Back to diagram overview

Compare with GGP-14-001 · Label the observations directly.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(station=rep(c("North","South"),each=4),
                day=rep(1:4,2), reading=c(8,12,11,15,13,10,14,12))
ends <- d[d$day==max(d$day),]
p <- ggplot(d,aes(day,reading,colour=station,group=station)) +
  geom_line() + geom_text(data=ends,aes(label=station),hjust=0,nudge_x=.15) +
  coord_cartesian(xlim=c(1,5.2)) + theme(legend.position="none")
export_plot(p, "GGP-14-002")
ParameterMeaning in this example
data=endsUse a deliberately selected row subset for the label layer.

Render this diagram

tools/render examples/GGP-14-002.R

Try changing: Move the labels by changing nudge_x to .3.

Actual output

Name each series at its end
GGP-14-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-14-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-14-003

Omit colliding labels deliberately

Check-overlap can suppress labels that would collide while retaining all points. State this behavior when an inventory of every identifier matters.

Back to diagram overview

Compare with GGP-14-002 · Name each series at its end.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(x=c(1,1.03,1.06,2,3),y=c(2,2.02,2.04,3,4),
                id=c("R17","R18","R19","R20","R21"))
p <- ggplot(d,aes(x,y)) + geom_point() +
  geom_text(aes(label=id),nudge_y=.15,check_overlap=TRUE) +
  scale_y_continuous(expand=expansion(mult=c(.1,.2))) +
  labs(caption="Some labels may be omitted; all points remain")
export_plot(p, "GGP-14-003")
ParameterMeaning in this example
check_overlap=TRUESuppress colliding text without filtering observations.

Render this diagram

tools/render examples/GGP-14-003.R

Try changing: Spread the first three positions farther apart.

Actual output

Omit colliding labels deliberately
GGP-14-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-14-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-14-004

Give a label its own background

A label geometry adds a background around text. Use it for a small set of important records when a plain text label would blend into the panel.

Back to diagram overview

Compare with GGP-14-003 · Omit colliding labels deliberately.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(x=c(1,3,5),y=c(2,4,3),id=c("R17","R18","R19"))
p <- ggplot(d,aes(x,y,label=id)) +
  geom_label(fill="#E8F1FA",linewidth=.25) +
  scale_y_continuous(expand=expansion(mult=.25))
export_plot(p, "GGP-14-004")
ParameterMeaning in this example
geom_label()Draw the text with a separate background and boundary.

Render this diagram

tools/render examples/GGP-14-004.R

Try changing: Change the background to white.

Actual output

Give a label its own background
GGP-14-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-14-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-14-005

Point to a stated event

An arrow annotation can identify one supplied observation. Keep its text away from the arrow tip and describe the event without claiming that it caused the measured change.

Back to diagram overview

Compare with GGP-14-004 · Give a label its own background.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(day=1:5, temperature=c(12,14,13,16,15))
p <- ggplot(d,aes(day,temperature)) + geom_point(size=3) +
  annotate("segment",x=3,y=17,xend=3.9,yend=16.1,
           arrow=grid::arrow(length=grid::unit(3,"mm"))) +
  annotate("text",x=3,y=17.7,label="Inspection") +
  scale_y_continuous(expand=expansion(mult=c(.05,.2)))
export_plot(p, "GGP-14-005")
ParameterMeaning in this example
grid::arrow()Add a native physical-size arrowhead to the annotation.

Render this diagram

tools/render examples/GGP-14-005.R

Try changing: Move the arrow and note to another observation.

Actual output

Point to a stated event
GGP-14-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-14-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-14-006

Inset a second ggplot figure

A ggplot grob can be placed inside another panel. Reserve an unoccupied region so the inset does not cover the observations it is meant to help explain.

Back to diagram overview

Compare with GGP-14-005 · Point to a stated event.

Complete R / ggplot2 source

source("config/style.R")
d <- data.frame(x=1:6,y=c(8,10,9,13,14,13))
q <- ggplot(d,aes(y)) + geom_histogram(binwidth=2,fill="#00549F") +
  theme_minimal(base_size=7,base_family="Source Sans 3")
p <- ggplot(d,aes(x,y)) + geom_point(size=3) +
  annotation_custom(ggplotGrob(q),xmin=1.1,xmax=3,ymin=13,ymax=17) +
  coord_cartesian(ylim=c(7,18))
export_plot(p, "GGP-14-006")
ParameterMeaning in this example
ggplotGrob()Convert a native ggplot into an inset drawing object.

Render this diagram

tools/render examples/GGP-14-006.R

Try changing: Change the inset bin width to one.

Actual output

Inset a second ggplot figure
GGP-14-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-14-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

15 · Local data and validation

Read the bundled fixtures as data, check their assumptions, and make each transformation explicit.

GGP-15-001

Read the bundled observations

A CSV reader supplies the same complete local table to every figure. Use explicit columns and inspect the accompanying data dictionary before plotting.

Back to diagram overview

Compare with GGP-02-002 · Keep series in a long table.

Complete R / ggplot2 source

source("config/style.R")
d <- read.csv("data/measurements.csv",stringsAsFactors=FALSE)
p <- ggplot(d,aes(day,temperature,colour=station)) + geom_point(size=3)
export_plot(p, "GGP-15-001")
ParameterMeaning in this example
read.csv()Read the bundled file with its header and numeric columns.

Render this diagram

tools/render examples/GGP-15-001.R

Try changing: Edit one temperature in the local CSV.

Actual output

Read the bundled observations
GGP-15-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-15-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-15-002

Check the input schema before drawing

A validation statement can stop a chart from using an incompatible table. Use explicit column and finite-value checks when silently plotting partial data would mislead.

Back to diagram overview

Compare with GGP-15-001 · Read the bundled observations.

Complete R / ggplot2 source

source("config/style.R")
d <- read.csv("data/measurements.csv",stringsAsFactors=FALSE)
stopifnot(all(c("day","station","temperature") %in% names(d)),
          is.numeric(d$temperature),all(is.finite(d$temperature)))
p <- ggplot(d,aes(day,temperature,colour=station)) + geom_point(size=3)
export_plot(p, "GGP-15-002")
ParameterMeaning in this example
stopifnot()Reject the specified schema or invalid numeric values before plotting.

Render this diagram

tools/render examples/GGP-15-002.R

Try changing: Temporarily rename the temperature header and inspect the validation error.

Actual output

Check the input schema before drawing
GGP-15-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-15-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-15-003

Convert a stored measurement unit

A derived field can carry the display unit while preserving the input column. Use an explicit formula and matching axis label when reporting in a different unit.

Back to diagram overview

Compare with GGP-15-002 · Check the input schema before drawing.

Complete R / ggplot2 source

source("config/style.R")
d <- read.csv("data/measurements.csv",stringsAsFactors=FALSE)
d$fahrenheit <- d$temperature*9/5+32
p <- ggplot(d,aes(day,fahrenheit,colour=station)) + geom_point(size=3) +
  labs(y="Temperature (F)")
export_plot(p, "GGP-15-003")
ParameterMeaning in this example
temperature*9/5+32Convert the stored Celsius values to Fahrenheit.

Render this diagram

tools/render examples/GGP-15-003.R

Try changing: Plot the original Celsius column with its matching label.

Actual output

Convert a stored measurement unit
GGP-15-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-15-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-15-004

Aggregate the supplied station readings

A summary table can be calculated before building a plot. Here each station mean uses all six finite fixture observations, making the calculation independent of the drawing geometry.

Back to diagram overview

Compare with GGP-15-003 · Convert a stored measurement unit.

Complete R / ggplot2 source

source("config/style.R")
d <- read.csv("data/measurements.csv",stringsAsFactors=FALSE)
summary <- aggregate(temperature~station,data=d,FUN=mean)
p <- ggplot(summary,aes(station,temperature)) + geom_point(size=4) +
  labs(y="Mean temperature (C)")
export_plot(p, "GGP-15-004")
ParameterMeaning in this example
aggregate()Calculate one explicit summary row per station.

Render this diagram

tools/render examples/GGP-15-004.R

Try changing: Replace mean with median.

Actual output

Aggregate the supplied station readings
GGP-15-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-15-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-15-005

Reshape a wide input table

A reshape operation can turn separate station columns into a long plotting table. Use it when the original file format should stay convenient for data entry.

Back to diagram overview

Compare with GGP-15-004 · Aggregate the supplied station readings.

Complete R / ggplot2 source

source("config/style.R")
wide <- read.csv("data/station-counts.csv")
d <- reshape(wide,direction="long",varying=list(c("North","South")),
             v.names="count",timevar="station",times=c("North","South"),
             idvar="day")
p <- ggplot(d,aes(day,count,colour=station)) + geom_line() + geom_point()
export_plot(p, "GGP-15-005")
ParameterMeaning in this example
reshape(direction="long")Turn the two station columns into a group field and value field.

Render this diagram

tools/render examples/GGP-15-005.R

Try changing: Add another day to the wide CSV.

Actual output

Reshape a wide input table
GGP-15-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-15-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-15-006

Attach station-specific references

A reference layer can read its own local table. The shared station field places each supplied limit in the correct facet instead of repeating every limit everywhere.

Back to diagram overview

Compare with GGP-15-005 · Reshape a wide input table.

Complete R / ggplot2 source

source("config/style.R")
d <- read.csv("data/measurements.csv",stringsAsFactors=FALSE)
targets <- read.csv("data/station-targets.csv")
p <- ggplot(d,aes(day,temperature)) + geom_point(size=2) +
  geom_hline(data=targets,aes(yintercept=limit),linetype="dashed") +
  facet_wrap(vars(station)) + labs(y="Temperature (C)")
export_plot(p, "GGP-15-006")
ParameterMeaning in this example
data=targetsRead reference values from a separate table with matching station keys.

Render this diagram

tools/render examples/GGP-15-006.R

Try changing: Change the South limit in its local table.

Actual output

Attach station-specific references
GGP-15-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-15-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

16 · One archive in several views

Answer distinct questions with the same small fixture and keep every denominator and unit visible.

GGP-16-001

Report archive intake by station

A station total answers how many records entered the archive. Sum the daily counts before drawing and keep the zero baseline for a proportional count comparison.

Back to diagram overview

Compare with GGP-06-002 · Rank categories by their totals.

Complete R / ggplot2 source

source("config/style.R")
d <- read.csv("data/measurements.csv",stringsAsFactors=FALSE)
totals <- aggregate(records~station,data=d,FUN=sum)
p <- ggplot(totals,aes(station,records)) + geom_col(fill="#00549F") +
  labs(title="Archive intake",y="Records")
export_plot(p, "GGP-16-001")
ParameterMeaning in this example
sum by stationUse every supplied daily count in its station total.

Render this diagram

tools/render examples/GGP-16-001.R

Try changing: Change the aggregation to show each day instead.

Actual output

Report archive intake by station
GGP-16-001 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-16-001: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-16-002

Compare the room temperature trends

The same table also describes environmental conditions over time. Connect only observations from the same station and state the unit on the axis.

Back to diagram overview

Compare with GGP-16-001 · Report archive intake by station.

Complete R / ggplot2 source

source("config/style.R")
d <- read.csv("data/measurements.csv",stringsAsFactors=FALSE)
p <- ggplot(d,aes(day,temperature,colour=station,group=station)) +
  geom_line() + geom_point() + labs(title="Room conditions",y="Temperature (C)")
export_plot(p, "GGP-16-002")
ParameterMeaning in this example
group=stationKeep the two supplied room trends independent.

Render this diagram

tools/render examples/GGP-16-002.R

Try changing: Add the station limits as a separate reference layer.

Actual output

Compare the room temperature trends
GGP-16-002 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-16-002: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-16-003

Show observations beside their summaries

A distribution summary can retain all original observations. Use a fixed jitter seed so the display remains stable and readers can see the small sample size.

Back to diagram overview

Compare with GGP-16-002 · Compare the room temperature trends.

Complete R / ggplot2 source

source("config/style.R")
d <- read.csv("data/measurements.csv",stringsAsFactors=FALSE)
p <- ggplot(d,aes(station,temperature)) +
  geom_boxplot(width=.35,outlier.shape=NA,fill="#E8F1FA") +
  geom_point(position=position_jitter(width=.06,height=0,seed=17)) +
  labs(title="Temperature distributions",y="Temperature (C)")
export_plot(p, "GGP-16-003")
ParameterMeaning in this example
seed=17Keep the point displacement reproducible.
outlier.shape=NAAvoid drawing an observation twice; the point layer retains every row.

Render this diagram

tools/render examples/GGP-16-003.R

Try changing: Reduce the jitter width to .03.

Actual output

Show observations beside their summaries
GGP-16-003 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-16-003: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-16-004

Report release rates with their denominators

A release rate answers a different question from an intake count. Calculate each station's ratio from summed released and received counts, not an unweighted mean of daily percentages.

Back to diagram overview

Compare with GGP-16-003 · Show observations beside their summaries.

Complete R / ggplot2 source

source("config/style.R")
d <- read.csv("data/measurements.csv",stringsAsFactors=FALSE)
totals <- aggregate(cbind(records,released)~station,data=d,FUN=sum)
totals$percent <- 100*totals$released/totals$records
p <- ggplot(totals,aes(station,percent)) + geom_col(fill="#00549F") +
  scale_y_continuous(limits=c(0,100)) + labs(y="Released (%)")
export_plot(p, "GGP-16-004")
ParameterMeaning in this example
Ratio of sumsUse each station's full received count as its denominator.

Render this diagram

tools/render examples/GGP-16-004.R

Try changing: Inspect the totals table before plotting.

Actual output

Report release rates with their denominators
GGP-16-004 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-16-004: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-16-005

Scan station-days as a heatmap

A heatmap reveals the same temperature table as a matrix of station-days. Use a fixed value domain so the colors retain the same meaning in another reporting period.

Back to diagram overview

Compare with GGP-16-004 · Report release rates with their denominators.

Complete R / ggplot2 source

source("config/style.R")
d <- read.csv("data/measurements.csv",stringsAsFactors=FALSE)
p <- ggplot(d,aes(day,station,fill=temperature)) + geom_tile() +
  scale_fill_gradient(low="#E8F1FA",high="#00549F",limits=c(10,20)) +
  labs(title="Station-day conditions",fill="Temperature (C)")
export_plot(p, "GGP-16-005")
ParameterMeaning in this example
Fixed temperature domainKeep the color meaning stable across all station-days.

Render this diagram

tools/render examples/GGP-16-005.R

Try changing: Change the common domain to 8 through 22.

Actual output

Scan station-days as a heatmap
GGP-16-005 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-16-005: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

GGP-16-006

Compose an intake and conditions report

Two native ggplot objects can share one exported page. Use distinct titles and units so the count panel and temperature panel cannot be mistaken for the same quantity.

Back to diagram overview

Compare with GGP-16-005 · Scan station-days as a heatmap.

Complete R / ggplot2 source

source("config/style.R")
d <- read.csv("data/measurements.csv",stringsAsFactors=FALSE)
totals <- aggregate(records~day,data=d,FUN=sum)
a <- ggplot(totals,aes(day,records)) + geom_col(fill="#00549F") +
  labs(title="Intake",y="Records")
b <- ggplot(d,aes(day,temperature,colour=station,group=station)) +
  geom_line() + labs(title="Conditions",y="Temperature (C)")
plots <- matrix(list(ggplotGrob(a),ggplotGrob(b)),nrow=1)
p <- gtable::gtable_matrix("archive",plots,widths=grid::unit(c(1,1),"null"),
                           heights=grid::unit(1,"null"))
export_plot(p, "GGP-16-006")
ParameterMeaning in this example
gtable_matrix()Compose the two native plotting objects without a raster screenshot.

Render this diagram

tools/render examples/GGP-16-006.R

Try changing: Exchange the two panels in the matrix.

Actual output

Compose an intake and conditions report
GGP-16-006 · R / ggplot2 native output

R / ggplot2 · exit 0

Native compiler log

Standard output

GGP-16-006: native PDF, SVG and PNG rendered with R 4.6.2; ggplot2 4.0.4

Download: Source · Source + configuration ZIP · SVG · Native PDF · PNG

Choosing a diagram

Start with the question and the relationship to communicate. Use the smallest example that contains the needed idea, then add labels, grouping and styling only where they clarify it. Inspect the native figure at its intended reading size. The examples draw supplied process meanings and statistics; a drawing does not execute a workflow or establish a causal claim.

Goal Start here
Compare category totals or rates Chapter 06
Inspect a distribution Chapter 07
State uncertainty or a summary rule Chapter 08
Compare groups in panels Chapter 10
Read a calendar axis Chapter 12
Show a matrix or field Chapter 13
Compose a finished report Chapter 16

Feature coverage

Feature family Chapter
Build a plot 01
Data tables and groups 02
Layers and geometry 03
Scales and domains 04
Labels and themes 05
Bars and rates 06
Distributions 07
Uncertainty and summaries 08
Relationships and models 09
Facets and panels 10
Coordinates and reference geometry 11
Dates and time series 12
Matrices and fields 13
Annotation and composition 14
Local data and validation 15
One archive in several views 16

References and licensing

Official ggplot2 reference, R manuals and the installed package help for ggplot2 4.0.4.

Compare compact native chart programs in the gnuplot tour and array-based work in the GNU Octave tour. The data visualization comparison runs matched Titanic views across native tools, keeping every statistical definition fixed.

Original prose, programs and synthetic datasets are CC0. Copyrightable structure and the publishing framework are CC BY 4.0, Lukas C. Bossert. Bundled fonts, package sources and other third-party material retain their own terms; consult the license scope and LICENSES/. The shared example and command index links every rendered program below.

Example and command index

Content: CC0 1.0. Structure and publishing framework: CC BY 4.0 by Lukas C. Bossert. Repository source · Publishing framework source · License scope and third-party notices.