Package 'coursekata'

Description

Data describing all residential home sales in Ames, Iowa from the years 2006–2010 as reported by the Ames City Assessor's Office and compiled by De Cock (2011). Ames is located about 30 miles north of Des Moines (the stats capitol) and is home to Iowa State University (the largest university in the state). Each row represents the latest sale of a home (one row per home in the dataset). Columns represent home features and sale prices (outcome). The original dataset includes a uniquely detailed (81 features per home) and comprehensive look at the housing market. The data included here are only a subset used for examples in CourseKata course material. See the references and data source for the full dataset.

Pedagogical Modifications

To simplify the dataset for instructional purposes, the data were filtered to include only single family homes, residential zoning, 1-2 story homes, homes with brick, cinder block, or concrete foundations, and average to excellent kitchen qualities. Further, the descriptive variables were reduced to the subset described in the format section.

Usage

Ames

Format

A data frame with 2930 observations on the following 80 variables:

YearBuilt

Year home was built (YYYY).

YearSold

Year of home sale (YYYY). Note: all home sales in this dataset occurred between 2006 - 2010. If a home was sold more than once between 2006 - 2010, only its latest sale is included in dataset.

Neighborhood

One of two neighborhoods in Ames county:

• College Creek (CollegeCreek), a neighborhood located adjacent to Iowa State University (the largest University in the state).

• Old Town (OldTown), a nationally designated historic district in Ames. The old neighborhood is located just north of the central business district.

HomeSizeR

Raw above-ground area of home, measured in square feet.

HomeSizeK

Above-ground area of home, measured in thousands of square feet.

LotSizeR

Raw total property lot size, measured in square feet.

LotSizeK

Total property lot size, in thousands of square feet.

Floors

Number of above-ground floors (1 story or 2 story).

BuildQuality

Assessor's rating of overall material and finish of the house.

• 10: Very Excellent

• 9: Excellent

• 8: Very Good

• 7: Good

• 6: Above Average

• 5: Average

• 4: Below Average

• 3: Fair

• 2: Poor

• 1: Very Poor

Foundation

Type of foundation (ground material underneath the house).

• Brick&Tile: Brick and Tile

• CinderBlock: Cinder Blocks

• PouredConcrete: Poured Concrete

HasCentralAir

Indicator if home contains central air conditioning (0 = No, 1 = Yes).

Bathrooms

Number of full above-ground bathrooms.

Bedrooms

Number of full above-ground bedrooms.

TotalRooms

Number of above-ground rooms in home, excluding bathrooms.

KitchenQuality

Assessor's rating of kitchen material quality.

• Excellent

• Good

• Average

HasFireplace

Indicator if home contains at least one fireplace (0 = No, 1 = Yes).

GarageType

Type of garage.

• Attached: includes attached, built-in, basement, and dual-type garages

• Detached: includes detached and carport garages

• None: home does not have a garage or carport

GarageCars

Number of cars that can fit in garage.

PriceR

Sale price of home, in raw USD ($)

PriceK

Sale price of home, in thousands of USD ($)

TinySet

(Ignore) Whether or not this row is in ames_tiny.csv

Source

https://www.kaggle.com/competitions/house-prices-advanced-regression-techniques/data

References

De Cock, Dean, (2011). Ames, Iowa: Alternative to the Boston Housing Data as an end of semester regression project, Journal of Statistics Education, 19(3). doi:10.1080/10691898.2011.11889627

Description

These data were generated as outcomes for "students" for three different "instructors" named A, B, and C. The outcome have means such that C > B > A, but the difference is only clearly significant for C > A, and borderline for the others.

Usage

class_data

Format

An object of class tbl_df (inherits from tbl, data.frame) with 105 rows and 2 columns.

Details

outcome

A hypothetical, numerical outcome of an intervention.

teacher

Either "A", "B", or "C", associating the outcome to a teacher.

Description

Attach the CourseKata course packages

Usage

coursekata_attach(do_not_ask = FALSE, quietly = FALSE)

Arguments

Value

A named logical vector indicating which packages were attached.

Examples

coursekata_attach()

Description

Install or update all CourseKata packages.

Usage

coursekata_install(...)

coursekata_update(...)

Arguments

Value

The state of all the packages after any updates have been performed.

Description

This function is called at package start-up and should rarely be needed by the user. The exception is when the user has called coursekata_unload_theme() and wants to go back to the CourseKata look and feel. When run, this function sets the CourseKata color palettes coursekata_palette(), sets the default theme to theme_coursekata(), and tweaks some default settings for specific plots. To restore the plotting settings that were active before this function first changed them, run coursekata_unload_theme().

Usage

coursekata_load_theme()

Value

No return value, called to adjust the global state of ggplot2.

See Also

coursekata_palette theme_coursekata scale_discrete_coursekata coursekata_unload_theme

Description

List all CourseKata course packages

Usage

coursekata_packages(check_remote_version = FALSE)

Arguments

Value

A data frame with three variables: the name of the package package, the version, and whether it is currently attached.

Examples

coursekata_packages()

Description

The color palettes used in our theme system

Usage

coursekata_palette(indices = integer(0))

Arguments

Value

A named list of the requested colors in the palette.

Examples

coursekata_palette()
coursekata_palette(c(1, 3, 5))

Description

Create a function that provides a colorblind palette.

Usage

coursekata_palette_provider()

Value

A function that accepts one argument n, which is the number of colors you want to use in the plot. This function is used by scales like scale_color_discrete to provide colorblind- safe palettes. Where possible, the function will use the hand-picked colors from coursekata_palette(), and when more colors are needed than are available, it will use the viridisLite::viridis() palette.

See Also

scale_discrete_coursekata

Examples

palette <- coursekata_palette_provider()
palette(3)

Description

Ensures a default CRAN is set if one is not already set.

Usage

coursekata_repos(repos = getOption("repos"))

Arguments

Value

A set of repositories that can be used to install or update the CourseKata packages.

Examples

coursekata_repos()

Description

This function restores the theme, options, and geom defaults that were active before coursekata_load_theme() first changed them. Calling it again before another load has no effect. To go back to the CourseKata look and feel, run coursekata_load_theme().

Usage

coursekata_unload_theme()

Value

No return value, called to restore the global state of ggplot2.

See Also

coursekata_load_theme

Description

Data from: Controlled clinical trial of canine therapy versus usual care to reduce patient anxiety in the emergency department.

Abstract

Objective

Test if therapy dogs can reduce anxiety in emergency department (ED) patients.

Methods

In this controlled clinical trial (NCT03471429), medically stable, adult patients were approached if the physician believed that the patient had “moderate or greater anxiety.” Patients were allocated on a 1:1 ratio to either 15 min exposure to a certified therapy dog and handler (dog), or usual care (control). Patient reported anxiety, pain and depression were assessed using a 0-10 scale (10=worst). Primary outcome was change in anxiety from baseline (T0) to 30 min and 90 min after exposure to dog or control (T1 and T2 respectively); secondary outcomes were pain, depression and frequency of pain medication.

Results

Among 98 patients willing to participate in research, 7 had aversions to dogs, leaving 91 (93%) were willing to see a dog; 40 patients were allocated to each group (dog or control). No data were normally distributed. Median baseline anxiety, pain and depression were similar between groups. With dog exposure, anxiety decreased significantly from T0 to T1: 6 (IQR 4-9.75) to T1: 2 (0-6) compared with 6 (4-8) to 6 (2.5-8) in controls (P<0.001, for T1, Mann-Whitney U). Dog exposure was associated with significantly lower anxiety at T2 and a significant overall treatment effect on two-way repeated measures ANOVA for anxiety, pain and depression. After exposure, 1/40 in the dog group needed pain medication, versus 7/40 in controls (P=0.056, Fisher’s).

Conclusions

Exposure to therapy dogs plus handlers significantly reduced anxiety in ED patients.

Usage

er

Format

A data frame with 84 observations on the following 53 variables:

id

Subject ID

condition

Whether the subject saw a Dog or was in the Control group

age

Subject's age in years

gender

Subject's self-identified gender

race

Subject's self-identified race

veteran

Is the subject a veteran?

disabled

Is the subject disabled?

dog_name

The name of the therapy dog

base_pain

Subject's self reported pain before the intervention (T0)

base_depression

Subject's self reported depression before the intervention (T0)

base_anxiety

Subject's self reported anxiety before the intervention (T0)

base_total

The sum of the subject's ⁠base_*⁠ scores

later_pain

Subject's self reported pain after the intervention (T1)

later_depression

Subject's self reported depression after the intervention (T1)

later_anxiety

Subject's self reported anxiety after the intervention (T1)

later_total

The sum of the subject's ⁠later_*⁠ scores

last_pain

Subject's self reported pain after the intervention (T2)

last_depression

Subject's self reported depression after the intervention (T2)

last_anxiety

Subject's self reported anxiety after the intervention (T2)

last_total

The sum of the subject's ⁠last_*⁠ scores

change_pain

The change in subject's pain from before the intervention to after

change_depression

The change in subject's depression from before the intervention to after

change_anxiety

The change in subject's anxiety from before the intervention to after

change_total

The sum of the subject's ⁠change_*⁠ scores

provider_male

Was the health care provider male?

provider

The health care provider's status: either an ⁠Advanced Practitioner⁠, Resident physician, or Attending physician

heart_rate

The subject's heart rate at baseline (T0)

resp_rate

The subject's respiratory rate at baseline (T0)

sp_o2

The subject's SpO2 at baseline (T0)

bp_syst

The subject's systolic blood pressure at baseline (T0)

bp_diast

The subject's diastolic blood pressure at baseline (T0)

med_given

Was the subject given medication prior to the study? (T0)

mh_none

None of the other medical history items were indicated

mh_asthma

Medical history: asthma

mh_smoker

Medical history: smoker

mh_cad

Medical history: coronary artery disease

mh_diabetes

Medical history: diabetes mellitus

mh_hypertension

Medical history: hypertension

mh_stroke

Medical history: prior stroke

mh_chronic_kidney

Medical history: chronic kidney disease

mh_copd

Medical history: chronic obstructive pulmonary disease

mh_hyperlipidemia

Medical history: hyperlipidemia

mh_hiv

Medical history: HIV

mh_other

Medical history: other (write-in)

ph_adhd

Psychiatric history: attention-deficit/hyperactivity disorder

ph_anxiety

Psychiatric history: anxiety

ph_bipolar

Psychiatric history: bipolar

ph_borderline

Psychiatric history: borderline personality disorder

ph_depression

Psychiatric history: depression

ph_schizophrenia

Psychiatric history: schizophrenia

ph_ptsd

Psychiatric history: PTSD

ph_none

None of the other psychiatric history items were indicated

ph_other

Psychiatric history: other (write-in)

References

Kline, J. A., Fisher, M. A., Pettit, K. L., Linville, C. T., & Beck, A. M. (2019). Controlled clinical trial of canine therapy versus usual care to reduce patient anxiety in the emergency department. PloS One, 14(1), e0209232. doi:10.1371/journal.pone.0209232

Description

This collection of functions is useful for extracting estimates and statistics from a fitted model. They are particularly useful when estimating many models, like when bootstrapping confidence intervals. Each function can be used with an already fitted model as an lm object, or a formula and associated data can be passed to it. All of these assume the comparison is the empty model.

Usage

b0(object, data = NULL)

b1(object, data = NULL)

b(object, data = NULL, all = FALSE, predictor = character())

f(object, data = NULL, all = FALSE, predictor = character(), type = 3)

pre(object, data = NULL, all = FALSE, predictor = character(), type = 3)

p(object, data = NULL, all = FALSE, predictor = character(), type = 3)

Arguments

Details

• b0: The intercept from the full model.

• b1: The slope b1 from the full model.

• b: The coefficients from the full model.

• f: The F value from the full model.

• pre: The Proportional Reduction in Error for the full model.

• p: The p-value from the full model.

• sse: The SS Error (SS Residual) from the model.

• ssm: The SS Model (SS Regression) for the full model.

• ssr: Alias for SSM.

fVal() and PRE() are older names for f() and pre(). They are kept for backward compatibility and behave identically to the newer functions.

Value

The value of the estimate as a single number.

References

Judd, C. M., McClelland, G. H., & Ryan, C. S. (2017). Data Analysis: A Model Comparison Approach to Regression, ANOVA, and Beyond (3rd ed.). New York: Routledge. ISBN:879-1138819832

Examples

supernova(lm(mpg ~ disp, data = mtcars))

change_p_decimals <- supernova(lm(mpg ~ disp, data = mtcars))
print(change_p_decimals, pcut = 8)

Description

Data from: Fundamentals of Biostatistics Notes from: Kahn, M.

Abstract

Sample of 654 youths, aged 3 to 19, in the area of East Boston during middle to late 1970's. Interest concerns the relationship between smoking and FEV. Since the study is necessarily observational, statistical adjustment via regression models clarifies the relationship.

Pedagogical Notes:

This is a versatile dataset that can be used throughout an introductory statistics course as well as an introductory modeling course. It includes many issues from statistical adjustment in observational studies, to subgroup analysis, quadratic regression and analysis of covariance.

Usage

fevdata

Format

A data frame with 654 observations on the following 5 variables:

AGE

Age, in years

FEV

Forced expiratory volume, in liters

HEIGHT

Height, in inches

SEX

0 = Female, 1 = Male

SMOKE

0 = Non-smoker, 1 = Smoker

References

Kahn,M. (2003). Data Sleuth, STATS, 37, 24. https://jse.amstat.org/datasets/fev.txt Rosner, B. (1999). Fundamentals of Biostatistics, Pacific Grove, CA: Duxbury

Description

Students at a university taking an introductory statistics course were asked to complete this survey as part of their homework.

Usage

Fingers

Format

A data frame with 157 observations on the following 16 variables:

Gender

Gender of participant.

RaceEthnic

Racial or ethnic background.

FamilyMembers

Members of immediate family (excluding self).

SSLast

Last digit of social security number (NA if no SSN).

Year

Year in school: 1=First, 2=Second, 3=Third, 4=Fourth, 5=Other

Job

Current employment status: 1=Not Working, 2=Part-time Job, 3=Full-time Job

MathAnxious

Agreement with the statement "In general I tend to feel very anxious about mathematics": 1=Strongly Disagree, 2=Disagree, 3=Neither Agree nor Disagree, 4=Agree, 5=Strongly Agree

Interest

Interest in statistics and the course: 1=No Interest, 2=Somewhat Interested, 3=Very Interested

GradePredict

Numeric prediction for final grade in the course. The value is converted from the student's letter grade prediction. 4.0=A, 3.7=A-, 3.3=B+, 3.0=B, 2.7=B-, 2.3=C+, 2.0=C, 1.7=C-, 1.3=Below C-

Thumb

Length in mm from tip of thumb to the crease between the thumb and palm.

Index

Length in mm from tip of index finger to the crease between the index finger and palm.

Middle

Length in mm from tip of middle finger to the crease between the middle finger and palm.

Ring

Length in mm from tip of ring finger to the crease between the middle finger and palm.

Pinkie

Length in mm from tip of pinkie finger to the crease between the pinkie finger and palm.

Height

Height in inches.

Weight

Weight in pounds.

Sex

Sex of participant.

Description

This is the Fingers dataset before it was cleaned. In the cleaning process, we converted the values from numbers to appropriate types (where applicable), removed outliers that suggested data was input incorrectly, and we removed incomplete cases. The description for the dataset is: Students at a university taking an introductory statistics course were asked to complete this survey as part of their homework. (This is the same data set as the Fingers data)

Usage

FingersMessy

Format

A data frame with 157 observations on the following 16 variables:

Gender

Gender of participant.

RaceEthnic

Racial or ethnic background.

FamilyMembers

Members of immediate family (excluding self).

SSLast

Last digit of social security number (NA if no SSN).

Year

Year in school: 1=First, 2=Second, 3=Third, 4=Fourth, 5=Other

Job

Current employment status: 1=Not Working, 2=Part-time Job, 3=Full-time Job

MathAnxious

Agreement with the statement "In general I tend to feel very anxious about mathematics": 1=Strongly Disagree, 2=Disagree, 3=Neither Agree nor Disagree, 4=Agree, 5=Strongly Agree

Interest

Interest in statistics and the course: 1=No Interest, 2=Somewhat Interested, 3=Very Interested

GradePredict

Numeric prediction for final grade in the course. The value is converted from the student's letter grade prediction. 4.0=A, 3.7=A-, 3.3=B+, 3.0=B, 2.7=B-, 2.3=C+, 2.0=C, 1.7=C-, 1.3=Below C-

Thumb

Length in mm from tip of thumb to the crease between the thumb and palm.

Index

Length in mm from tip of index finger to the crease between the index finger and palm.

Middle

Length in mm from tip of middle finger to the crease between the middle finger and palm.

Ring

Length in mm from tip of ring finger to the crease between the middle finger and palm.

Pinkie

Length in mm from tip of pinkie finger to the crease between the pinkie finger and palm.

Height

Height in inches.

Weight

Weight in pounds.

Sex

Sex of participant.

Description

Test the fit of a model on a train and test set.

Usage

fit_stats(model, df_train, df_test)

fitstats(model, df_train, df_test)

Arguments

Value

A data frame with the fit statistics.

Examples

set.seed(123)
parts <- split_data(Fingers)
model <- lm(Thumb ~ Height, data = parts$train)
fit_stats(model, parts$train, parts$test)
fitstats(model, parts$train, parts$test)

Description

The simulated results of a small study comparing the effectiveness of three different computer- based math games in a sample of 105 fifth-grade students. All three games focused on the same topic and had identical learning goals, and none of the students had any prior knowledge of the topic.

Usage

game_data

Format

A data frame with 105 observations on the following 2 variables:

game

The game the student was randomly assigned to, coded as "A", "B", or "C".

outcome

Each student's score on the outcome test.

Description

Add an intercept and a rise-over-run triangle, or arrows for group differences. Marks use tagged segment, point, and text layers. See gf_b() for the mark geometry and appearance controls.

Usage

geom_b(
  mapping = NULL,
  data = NULL,
  ...,
  model = NULL,
  color = "#b599ed",
  label_color = "black",
  label_size = 3.5,
  arrow_linewidth = 0.5,
  show_b0 = TRUE,
  run = NULL,
  run_x = NULL,
  b0_alpha = 0.3,
  b0_linewidth = 0.8,
  b0_size = 4,
  arrow_nudge = 0.18,
  label_nudge = 0.08,
  inherit.aes = TRUE
)

Arguments

Details

This constructor returns a composite annotation. On addition it selects the observation source and model layer; its component stats calculate the marks when the plot builds. The model stat supplies its prediction grid once per build. Adding the annotation does not evaluate data callbacks or fit a model. There is no separate stat_b() entry point because an ordinary stat cannot select a sibling model layer or compose the different drawing geoms.

A supplied model keeps its fitted coefficients. With no model or local source, the annotation uses the plotted model, or computes an implied model through StatModel when no model layer is present. An explicit mapping or data selects the annotation's own source instead of a sibling model. Scale transformations and hard limits then affect the inferred fit just as they affect geom_model(). run and run_x use those transformed units for an inferred model. Scales and limits can be added before or after the annotation. Coordinate zooming does not change the fitted rows. Replacing plot data updates inferred coefficients and annotation placement; a supplied fit keeps its coefficients.

Inference requires one model in one panel. Pass a fitted model to annotate a faceted plot. A nonlinear curve has no single slope to annotate. A plotted model must use the identity position: a positional adjustment moves its marks away from the fitted claim.

Value

A composite annotation to add to a ggplot.

Examples

fit <- lm(Thumb ~ Height, data = Fingers)
ggplot2::ggplot(Fingers, ggplot2::aes(Height, Thumb)) +
  ggplot2::geom_point() + geom_model(model = fit) + geom_b(model = fit)

Description

geom_model() adds a model to an ordinary ggplot2::ggplot(). A supplied model may be a fit from stats::lm() or stats::aov(), or a two-sided formula that is fitted once against the layer data. With no model, the layer draws the model implied by its mapped positions, separately in each panel and group.

Usage

geom_model(
  mapping = NULL,
  data = NULL,
  stat = "model",
  position = "identity",
  ...,
  model = NULL,
  orientation = NA,
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE
)

stat_model(
  mapping = NULL,
  data = NULL,
  geom = "model",
  position = "identity",
  ...,
  model = NULL,
  orientation = NA,
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE
)

Arguments

Details

Empty models draw an intercept, numeric predictors draw fitted lines, and categorical predictors draw one short mark per group. A model with one numeric and one categorical predictor draws one line per category.

A supplied model is one fixed claim evaluated on a prediction grid built from the layer data.

The displayed predictor is read from the plot mapping or this layer's local mapping. One additional predictor may be present in the data: categories produce separate traces, and numeric values use the mean and mean plus or minus one standard deviation. Outcome-axis expressions such as log(y) are applied to the predictions. A supplied model is prepared once when added to the plot, including when data is a function or formula.

Override unrelated inherited aesthetics locally, or set inherit.aes = FALSE and supply a predictor mapping such as aes(x = Height). Without explicit data or positions, the layer follows the first observation layer, preferring points, so a model describes the rows and axes the plot actually shows.

Value

A ggplot2 layer.

Examples

fit <- lm(Thumb ~ Height, data = Fingers)
ggplot2::ggplot(Fingers, ggplot2::aes(Height, Thumb)) +
  ggplot2::geom_point() +
  geom_model(model = fit)

group_fit <- lm(Thumb ~ Sex, data = Fingers)
ggplot2::ggplot(Fingers, ggplot2::aes(Sex, Thumb)) +
  ggplot2::geom_jitter(width = 0.1) +
  geom_model(model = group_fit)

Description

These layers measure a fitted model directly from an ordinary ggplot2::ggplot(). A residual segment runs from the model's prediction to the observed value; an arrow at its last end points to the observation. A reduction segment runs from the prediction to the model's grand mean. The square variants draw areas proportional to those squared distances at a shared aspect ratio.

Usage

geom_resid(
  mapping = NULL,
  data = NULL,
  stat = "resid",
  position = "identity",
  ...,
  model = NULL,
  fun = NULL,
  orientation = NA,
  linewidth = 0.2,
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE
)

geom_square_resid(
  mapping = NULL,
  data = NULL,
  stat = "resid",
  position = "identity",
  ...,
  model = NULL,
  fun = NULL,
  orientation = NA,
  aspect = 4/6,
  alpha = 0.1,
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE
)

geom_reduce(
  mapping = NULL,
  data = NULL,
  stat = "reduce",
  position = "identity",
  ...,
  model = NULL,
  fun = NULL,
  orientation = NA,
  linewidth = 0.2,
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE
)

geom_square_reduce(
  mapping = NULL,
  data = NULL,
  stat = "reduce",
  position = "identity",
  ...,
  model = NULL,
  fun = NULL,
  orientation = NA,
  aspect = 4/6,
  alpha = 0.1,
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE
)

stat_resid(
  mapping = NULL,
  data = NULL,
  geom = "resid",
  position = "identity",
  ...,
  model = NULL,
  fun = NULL,
  orientation = NA,
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE
)

stat_reduce(
  mapping = NULL,
  data = NULL,
  geom = "resid",
  position = "identity",
  ...,
  model = NULL,
  fun = NULL,
  orientation = NA,
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE
)

Arguments

Details

When the observations have their own data or mappings, these layers follow the first point layer (or the first non-annotation layer when there are no points). Explicit layer data and mapping arguments take precedence.

With orientation = NA, a model's outcome determines the direction. The outcome expression must match an axis exactly: log(y) is refused for a model of y, because that distance is not the model's residual. An explicit orientation = "x" puts the outcome on y; "y" puts it on x and must agree with that mapping. Coordinate systems such as ggplot2::coord_flip() are applied later and do not change this argument.

A jittered point layer and its model layer must use the same ggplot2::position_jitter() object with a numeric seed. The model layer keeps the fitted endpoint fixed while moving the observed endpoint by the same amount as its point.

Positional expressions are evaluated reproducibly on the current data, so random mappings give the observations and residuals the same coordinates. Native layers still respond to later data and mapping changes. A new unseeded random mapping must be added before the residual layer, or carry its own fixed seed. Checks that need the rows or mappings run when the plot is built: axes, prediction, outcome axis, then reduction eligibility.

Native square layers inherit mapped aesthetics by default, including colour. gf_square_resid() and gf_square_reduce() default to inherit = FALSE for neutral square outlines. Set their inherit = TRUE to match this API.

Reduction layers express the ordinary least-squares sum-of-squares identity. They require an unweighted model with an intercept, no offset, and its stored model frame. On the fitted observations, the identity holds across the sums of the square areas, not separately for each observation or for new prediction data.

Value

A ggplot2 layer.

Examples

model <- lm(Thumb ~ Height, data = Fingers)
ggplot2::ggplot(Fingers, ggplot2::aes(Height, Thumb)) +
  ggplot2::geom_point() +
  geom_resid(model = model, colour = "firebrick")

ggplot2::ggplot(Fingers, ggplot2::aes(Height, Thumb)) +
  ggplot2::geom_point() +
  geom_square_reduce(model = model, fill = "forestgreen")

jitter <- ggplot2::position_jitter(width = 0.1, seed = 42)
group_model <- lm(Thumb ~ Sex, data = Fingers)
ggplot2::ggplot(Fingers, ggplot2::aes(Sex, Thumb)) +
  ggplot2::geom_point(position = jitter) +
  geom_resid(model = group_model, position = jitter)

Description

geom_squareplot() and stat_squareplot() expose gf_squareplot() through ggplot2 syntax. With the default bars = "none", each observation is drawn as one square. Continuous x values use ggplot2::stat_bin()'s binning rules; factors, characters, and logicals use ggplot2::stat_count()'s columns. Both constructors retain gf_squareplot()'s defaults, warnings, factor-level handling, and count scale. Counts must be finite, nonnegative whole numbers. Integer frequency weights are supported; use a histogram or bar plot for fractional weighted counts. Fill and separator colour inherit mappings, with teal and white defaults when no mapping or fixed value is supplied. Put a staged linewidth = after_scale(...) mapping on this layer to keep its width fixed. Plot-level after_scale() linewidth mappings use automatic border fitting: ggplot2 evaluates them after the geom's setup.

Usage

geom_squareplot(
  mapping = NULL,
  data = NULL,
  stat = "squareplot",
  position = "identity",
  ...,
  binwidth = NULL,
  bins = NULL,
  center = NULL,
  boundary = NULL,
  closed = NULL,
  breaks = NULL,
  bars = "none",
  na.rm = TRUE,
  show.legend = NA,
  inherit.aes = TRUE
)

stat_squareplot(
  mapping = NULL,
  data = NULL,
  geom = "squareplot",
  position = "identity",
  ...,
  binwidth = NULL,
  bins = NULL,
  center = NULL,
  boundary = NULL,
  closed = NULL,
  breaks = NULL,
  bars = "none",
  na.rm = TRUE,
  show.legend = NA,
  inherit.aes = TRUE
)

Arguments

Details

A squareplot is x-only, as it is through gf_squareplot(). Map the distribution to x. If the plot already maps y for another layer, set inherit.aes = FALSE and supply this layer's x mapping directly.

Value

A ggplot2 layer.

See Also

gf_squareplot()

Examples

ggplot2::ggplot(Fingers, ggplot2::aes(Thumb)) +
  geom_squareplot()

ratings <- data.frame(rating = factor(c(1, 1, 3), levels = 1:5))
ggplot2::ggplot(ratings, ggplot2::aes(rating)) +
  stat_squareplot()

Description

GeomCutoff draws a cutoff at each supplied xintercept. The statistical position is transformed by the active coordinate system, while height is measured as a fraction of the panel. The result is a short stem that starts at the distribution baseline and follows ggplot2::coord_flip() without changing the orthogonal scale. Optional label, side, and call_id aesthetics let a teaching helper add measured callouts without training the count axis. When the x scale is positioned at the top, the stem starts at that panel edge and points inward instead.

Usage

GeomCutoff

geom_cutoff(
  mapping = NULL,
  data = NULL,
  stat = "identity",
  position = "identity",
  ...,
  height = 0.2,
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE
)

Arguments

Format

GeomCutoff is a ggplot2::Geom object.

Details

geom_cutoff() is the conventional layer constructor. It does not compute cutoffs. show_cutoffs() supplies the optional marker and callout metadata after it computes one whole-distribution plan.

Value

geom_cutoff() returns a ggplot2 layer.

Description

GeomModel is a narrow dispatcher. The statistical representation chooses a line, group mark, or intercept; the corresponding ggplot2 geom does the drawing.

Usage

GeomModel

Format

GeomModel is a ggplot2::Geom object.

Description

The segment is drawn from what the model predicts to what was observed, so x/y is the observation and xend/yend the prediction until the moment of drawing. Which of the two ends arrives says which axis the residual is measured on. geom_resid() pairs it with StatResid and computes those endpoints from a fitted model.

Usage

GeomResid

Format

A ggplot2::Geom object.

See Also

gf_resid(), the ggformula function for the same layer.

Description

Re-expands each bin from StatSquareplot into one unit rectangle per observation, stacked. The rectangles are one unit tall in data units, so the y axis is a real count. Pair it with StatSquareplot in a ggplot2::layer() to draw the squares yourself, where a mapped fill stacks its groups within each bin.

Usage

GeomSquareplot

Format

A ggplot2::Geom object.

Details

The geom takes either a bin's edges (xmin/xmax, from ggplot2::stat_bin()) or a level's position plus a column width (x/width, from ggplot2::stat_count()), deriving the edges it was not given. Paired with stat = "count", it draws one countable column per category the way geom_bar() draws one bar.

The bars parameter chooses what a bin is drawn as, on one grid and in one layer: "none" draws the squares, "outline" frames them with the bar they add up to, and "solid" draws that bar with the squares covered over. The bar is derived from the squares rather than re-binned, so a bin holding no observations has no bar, and a solid bar stacks a mapped fill in the same spans its squares did. bar_color and bar_linewidth are the bar's own color and width; color and linewidth are the separators between the squares, and either may be mapped.

See Also

gf_squareplot(), which pairs this stat and geom for you.

Description

The square is expanded here rather than in StatResid because a stat runs before the position: four corners jittered one row at a time tear apart. It is also the only place the panel's final ranges are known, and the square is a square on the page rather than in data units.

Usage

GeomSquareResid

Format

A ggplot2::Geom object.

See Also

geom_square_resid() and gf_square_resid().

Description

Draws the intercept and slope (or group differences) of a fitted model as arrows and labels directly on the plot they describe: a rise-over-run triangle for a continuous predictor, an arrow to each non-reference group for a categorical one. Where gf_model() draws the fit itself, gf_b() draws the numbers that describe it.

gf_coef() is a fully supported alias of gf_b(). The package already exports b(), b0(), b1() as its vocabulary for coefficients, and a reader who knows stats::coef() will look for a plot-side counterpart under that name.

Usage

gf_b(
  object = NULL,
  gformula = NULL,
  data = NULL,
  ...,
  model,
  color = "#b599ed",
  label_color = "black",
  label_size = 3.5,
  arrow_linewidth = 0.5,
  show_b0 = TRUE,
  run = NULL,
  run_x = NULL,
  b0_alpha = 0.3,
  b0_linewidth = 0.8,
  b0_size = 4,
  arrow_nudge = 0.18,
  label_nudge = 0.08,
  xlab,
  ylab,
  title,
  subtitle,
  caption,
  geom = ggplot2::GeomSegment,
  stat = "identity",
  position = "identity",
  show.legend = NA,
  show.help = NULL,
  inherit = TRUE,
  environment = parent.frame()
)

gf_coef(
  object = NULL,
  gformula = NULL,
  data = NULL,
  ...,
  model,
  color = "#b599ed",
  label_color = "black",
  label_size = 3.5,
  arrow_linewidth = 0.5,
  show_b0 = TRUE,
  run = NULL,
  run_x = NULL,
  b0_alpha = 0.3,
  b0_linewidth = 0.8,
  b0_size = 4,
  arrow_nudge = 0.18,
  label_nudge = 0.08,
  xlab,
  ylab,
  title,
  subtitle,
  caption,
  geom = ggplot2::GeomSegment,
  stat = "identity",
  position = "identity",
  show.legend = NA,
  show.help = NULL,
  inherit = TRUE,
  environment = parent.frame()
)

Arguments

Value

A ggplot object with the model's coefficients annotated on it.

What is drawn

A continuous predictor: a vertical rise arrow from fit(run_x) to fit(run_x + run), a horizontal run segment at its tip, a rise label (plotmath b1 when run is 1, otherwise run times b1), a run-distance label on the triangle's interior side of the run segment, and a hollow dot at ⁠(0, b0)⁠ with a b0 label.

A categorical predictor: one horizontal reference line at b0 (the reference level's mean), and for each level after it a segment from b0 to b0 plus that level's coefficient, with an arrow head, labeled (plotmath) b1, b2, … Level order is read off coef(model), so a releveled factor still labels the arrow that matches its coefficient.

No predictor (the empty model): the b0 line and its label, nothing else.

Layers have stable role tags such as "b0", "b1", "run", and their "_label" counterparts. Categorical annotations use numbered roles plus shared "bk_more" and "bk_more_label" layers for additional groups. Replacement data can add groups without changing the layer list. A role can be empty when the model does not need it.

No model

With no model, gf_b() reads the model the plot implies – the same decision gf_model() uses. The marks read that model's prediction grid, after position scales transform values and remove rows outside hard limits. If no model layer exists, the same model stat supplies the predictions. Inference is refused on a faceted plot because its panels have different fits. Position scales and gf_lims() can precede or follow the annotation. For an inferred model on transformed scales, run and run_x use transformed units. geom_b() provides the same annotation with ggplot2 syntax.

Placement

Every mark is placed from the model's coefficients and from level indices, never from a drawn point's position, so jitter never moves an arrow.

show_b0 = TRUE (the default) expands the PREDICTOR's axis to include 0 on a continuous model, because b0 is the prediction where the predictor is 0 and a picture of it that does not show that point is not a picture of b0. Usually that is x; on a plot that puts the outcome on x it is y, and the expansion follows the predictor rather than the letter. Calling gf_lims() on that axis afterward overrides the expansion and can push the b0 dot off the page.

See Also

gf_model() draws the fit itself.

Examples

# continuous: b1 as a rise-over-run triangle, b0 where the line meets x = 0
height_model <- lm(Thumb ~ Height, data = Fingers)
gf_point(Thumb ~ Height, data = Fingers, alpha = .3) %>% gf_b(height_model)

# the slope per one unit
gf_point(Thumb ~ Height, data = Fingers) %>% gf_b(height_model, run = 1)

# an explicit run labels the rise "10 x b1"
gf_point(Thumb ~ Height, data = Fingers) %>% gf_b(height_model, run = 10)

# categorical: b0 is the reference group's mean, each b_k is an arrow to group k
tip_model <- lm(Tip ~ Condition, data = TipExperiment)
gf_jitter(Tip ~ Condition, data = TipExperiment, width = .1) %>% gf_b(tip_model)

# no model: the model the plot implies, on the values the plot drew
set.seed(1)
gf_jitter(shuffle(Height) ~ Sex, data = Fingers, width = .1) %>%
  gf_model() %>%
  gf_b()

# gf_coef() is the same function under the name coef() readers look for
flipper_model <- lm(body_mass_kg ~ flipper_length_m, data = penguins)
gf_point(body_mass_kg ~ flipper_length_m, data = penguins) %>%
  gf_coef(flipper_model)

Description

When teaching about regression it can be useful to visualize the data as a point plot with the outcome on the y-axis and the explanatory variable on the x-axis. For regression models, this is most easily achieved by calling ggformula::gf_lm(), with empty models ggformula::gf_hline() using the mean, and a more complicated call to ggformula::gf_segment() for group models. This function simplifies this by making a guess about what kind of model you are plotting (empty/null, regression, group) and then making the appropriate plot layer for it.

Usage

gf_model(
  object = NULL,
  gformula = NULL,
  data = NULL,
  ...,
  model,
  xlab,
  ylab,
  title,
  subtitle,
  caption,
  geom = "model",
  stat = "model",
  position = "identity",
  show.legend = NA,
  show.help = NULL,
  inherit = TRUE,
  environment = parent.frame()
)

Arguments

Details

This function only works with models that have a continuous outcome measure.

Value

A ggplot object with the model added. With no model, and a plot whose positional mapping is an expression rather than a bare variable (shuffle(body_mass_kg), log(flipper_length_m)), the RETURNED plot is pinned to the values it drew when gf_model() was called – its data gains a fixed column and its mapping names it, while its axis titles and everything else about how it reads keep your own words. The plot passed IN is untouched.

Supported plots

gf_model() is built for and tested against plots made with ggformula::gf_point(), ggformula::gf_jitter(), ggformula::gf_boxplot(), ggformula::gf_violin() and ggformula::gf_histogram(). Other plots may work if they map their variables the same way, but they are not tested.

Examples

# the empty model predicts the same value (the mean) for every observation
empty_model <- lm(body_mass_kg ~ NULL, data = penguins)
gf_histogram(~body_mass_kg, data = penguins, binwidth = 0.25) %>%
  gf_model(empty_model)

# a two-group model (categorical explanatory variable) on a jitter plot
gentoo_model <- lm(body_mass_kg ~ gentoo, data = penguins)
gf_jitter(body_mass_kg ~ gentoo, data = penguins, width = .1) %>%
  gf_model(gentoo_model)

# a three-group model works the same way
species_model <- lm(body_mass_kg ~ species, data = penguins)
gf_jitter(body_mass_kg ~ species, data = penguins, width = .1) %>%
  gf_model(species_model)

# group models can also be layered onto faceted histograms
gf_histogram(~body_mass_kg, data = penguins, binwidth = 0.25) %>%
  gf_facet_grid(species ~ .) %>%
  gf_model(species_model)

# a regression model (quantitative explanatory variable) on a scatter plot
flipper_model <- lm(body_mass_kg ~ flipper_length_m, data = penguins)
gf_point(body_mass_kg ~ flipper_length_m, data = penguins) %>%
  gf_model(flipper_model)

# layer the empty model and the regression model in different colors to
# compare the two models on the same plot
gf_point(body_mass_kg ~ flipper_length_m, data = penguins) %>%
  gf_model(empty_model, color = "dodgerblue") %>%
  gf_model(flipper_model, color = "firebrick")

# with a categorical and a quantitative predictor, the model is drawn
# as one line for each group
ancova_model <- lm(body_mass_kg ~ species + flipper_length_m, data = penguins)
gf_point(body_mass_kg ~ flipper_length_m, color = ~species, data = penguins) %>%
  gf_model(ancova_model)

# a model that has not been fit yet can be written as a formula, and is fit
# against the data the plot was built from
gf_point(body_mass_kg ~ flipper_length_m, data = penguins) %>%
  gf_model(body_mass_kg ~ flipper_length_m)

# the empty model, written as a formula
gf_histogram(~body_mass_kg, data = penguins, binwidth = 0.25) %>%
  gf_model(body_mass_kg ~ NULL)

# with no model, gf_model() draws the model the plot implies: a numeric
# predictor draws the regression line gf_lm() would fit
gf_point(body_mass_kg ~ flipper_length_m, data = penguins) %>%
  gf_model()

# a categorical predictor draws one mark at each group's mean
gf_jitter(body_mass_kg ~ species, data = penguins, width = .1) %>%
  gf_model()

# a plot that draws only its outcome implies the grand mean
gf_histogram(~body_mass_kg, data = penguins, binwidth = 0.25) %>%
  gf_model()

Description

Draws reduction lines from the value a fitted model predicts for each observation to the grand mean. Squaring and summing those lengths across observations gives the model sum of squares; gf_resid() supplies the error term in the same decomposition. Each line runs along whichever axis the plot puts the model's outcome on, so a model of the variable drawn on x is measured across x rather than down y.

Usage

gf_reduce(
  object = NULL,
  gformula = NULL,
  data = NULL,
  ...,
  model,
  linewidth = 0.2,
  xlab,
  ylab,
  title,
  subtitle,
  caption,
  geom = coursekata::GeomResid,
  stat = coursekata::StatReduce,
  position = "identity",
  show.legend = NA,
  show.help = NULL,
  inherit = TRUE,
  environment = parent.frame()
)

Arguments

Details

The grand mean is the model's own, mean() of the outcome column the model was fit on, not anything read off the plot's data. On a faceted plot that is one number for every panel: every panel is measured against the same line, which is what makes the picture in each panel a piece of one decomposition rather than a decomposition of its own.

Value

A ggplot object with reduction lines added.

Examples

set.seed(1)
penguins_20 <- sample(penguins, 20)

# the reduction: how far a model's fit moves the prediction from the grand
# mean, for a regression model
flipper_model <- lm(body_mass_kg ~ flipper_length_m, data = penguins_20)
gf_point(body_mass_kg ~ flipper_length_m, data = penguins_20) %>%
  gf_model(flipper_model) %>%
  gf_reduce(flipper_model, color = "blue")

# and for a two-group model on a jitter plot
gentoo_model <- lm(body_mass_kg ~ gentoo, data = penguins_20)
gf_jitter(body_mass_kg ~ gentoo, data = penguins_20, width = .1) %>%
  gf_model(gentoo_model) %>%
  gf_reduce(gentoo_model, color = "blue")

# each observation's signed deviation from the grand mean is its residual
# (firebrick) plus its reduction (blue)
gf_point(body_mass_kg ~ flipper_length_m, data = penguins_20) %>%
  gf_model(flipper_model) %>%
  gf_resid(flipper_model, color = "firebrick") %>%
  gf_reduce(flipper_model, color = "blue")

Description

Draws residual lines from the values a fitted model predicts to the observed points. Each residual runs along whichever axis the plot puts the model's outcome on, so a model of the variable drawn on x is measured across x rather than down y.

Usage

gf_resid(
  object = NULL,
  gformula = NULL,
  data = NULL,
  ...,
  model,
  linewidth = 0.2,
  xlab,
  ylab,
  title,
  subtitle,
  caption,
  geom = coursekata::GeomResid,
  stat = coursekata::StatResid,
  position = "identity",
  show.legend = NA,
  show.help = NULL,
  inherit = TRUE,
  environment = parent.frame()
)

Arguments

Value

A ggplot object with residual lines added.

Examples

# residuals can be drawn on a full data set, but with hundreds of points
# the plot gets hard to read
flipper_model <- lm(body_mass_kg ~ flipper_length_m, data = penguins)
gf_point(body_mass_kg ~ flipper_length_m, data = penguins) %>%
  gf_model(flipper_model) %>%
  gf_resid(flipper_model)

# a small sample makes the residuals much easier to see
set.seed(1)
penguins_20 <- sample(penguins, 20)

# residuals from the empty model (in blue)
empty_model <- lm(body_mass_kg ~ NULL, data = penguins_20)
gf_point(body_mass_kg ~ flipper_length_m, data = penguins_20) %>%
  gf_model(empty_model) %>%
  gf_resid(empty_model, color = "blue")

# residuals from a two-group model on a jitter plot (in firebrick)
gentoo_model <- lm(body_mass_kg ~ gentoo, data = penguins_20)
gf_jitter(body_mass_kg ~ gentoo, data = penguins_20, width = .1) %>%
  gf_model(gentoo_model) %>%
  gf_resid(gentoo_model, color = "firebrick")

# residuals from a regression model (in firebrick)
sample_flipper_model <- lm(body_mass_kg ~ flipper_length_m, data = penguins_20)
gf_point(body_mass_kg ~ flipper_length_m, data = penguins_20) %>%
  gf_model(sample_flipper_model) %>%
  gf_resid(sample_flipper_model, color = "firebrick")

Description

Usage

gf_resid_fun(
  object = NULL,
  gformula = NULL,
  data = NULL,
  ...,
  fun,
  linewidth = 0.2,
  xlab,
  ylab,
  title,
  subtitle,
  caption,
  geom = coursekata::GeomResid,
  stat = coursekata::StatResid,
  position = "identity",
  show.legend = NA,
  show.help = NULL,
  inherit = TRUE,
  environment = parent.frame()
)

Arguments

Details

Draws residual lines from predicted values to the observed points, using a user-supplied function of x (e.g., the function plotted with gf_function()). Where gf_resid() measures a fitted model, this measures a function you wrote: it is called on the x values the plot draws, and each residual runs from the prediction to the observation.

Value

A ggplot object with residual lines added.

Examples

set.seed(1)
df <- data.frame(X = 1:10, Y = 2 + 3 * (1:10) + rnorm(10))
my_fun <- function(x) 2 + 3 * x

gf_point(Y ~ X, data = df) %>%
  gf_function(my_fun) %>%
  gf_resid_fun(my_fun, color = "red", alpha = 0.5)

Description

Usage

gf_sd_ruler(
  object = NULL,
  gformula = NULL,
  data = NULL,
  ...,
  where = "middle",
  na.rm = FALSE,
  xlab,
  ylab,
  title,
  subtitle,
  caption,
  geom = "segment",
  stat = coursekata::StatSdRuler,
  position = "identity",
  show.legend = NA,
  show.help = NULL,
  inherit = TRUE,
  environment = parent.frame()
)

Arguments

Details

Adds a segment showing one standard deviation of the outcome, anchored at its mean. The orientation depends on where the outcome variable lives: on a scatter or jitter plot (outcome on the y-axis) the ruler is a vertical segment placed at a chosen x position; on a histogram (outcome on the x-axis, no y aesthetic) it is a horizontal segment running from the mean to mean + SD along the baseline. The orientation is detected automatically from the plot's axis mappings.

Both the outcome and, where relevant, the placement are measured in the space the panel is drawn in: a faceted plot measures each panel's own subset, and a transformed axis or a computed mapping such as ~log(Thumb) is measured in the transformed or computed values, not the raw column.

gf_sd_ruler() draws one ruler per panel, so an aesthetic mapped on the call – gf_sd_ruler(color = ~Sex) – is refused; split the plot instead with y ~ x | group to get one ruler per group.

Value

A ggplot object with the SD ruler segment added.

See Also

stat_sd_ruler() provides the ggplot2 interface. The model visualization guide shows the ruler alongside residuals and compares groups with different spread: https://coursekata.github.io/coursekata-r/articles/model-visualization.html

Examples

# the ruler runs from the mean (the empty model) up by one standard
# deviation -- it looks like a residual because SD is a typical residual
gf_point(Thumb ~ Height, data = Fingers, alpha = .4) %>%
  gf_model(lm(Thumb ~ NULL, data = Fingers)) %>%
  gf_sd_ruler()

# `where` controls placement along the x-axis
gf_point(Thumb ~ Height, data = Fingers, alpha = .4) %>%
  gf_sd_ruler(where = "mean")

# categorical x works the same way
gf_jitter(Thumb ~ Sex, data = Fingers, width = .1, alpha = .4) %>%
  gf_sd_ruler(where = "median")

# on a histogram the outcome is on the x-axis, so the ruler is horizontal
# and runs along the baseline from the mean to one SD above it
gf_histogram(~Thumb, data = Fingers, binwidth = 5) %>%
  gf_sd_ruler(linewidth = 2)

# name the variable explicitly when the plot does not make it obvious
gf_point(Thumb ~ Height, data = Fingers, alpha = .4) %>%
  gf_sd_ruler(Thumb ~ Height)

# one ruler per panel
gf_sd_ruler(Thumb ~ Height | Sex, data = Fingers)

Description

gf_squareduce() is a fully supported alias of gf_square_reduce(), named the way the classroom that asked for it says it.

Usage

gf_square_reduce(
  object = NULL,
  gformula = NULL,
  data = NULL,
  ...,
  model,
  aspect = 4/6,
  alpha = 0.1,
  xlab,
  ylab,
  title,
  subtitle,
  caption,
  geom = coursekata::GeomSquareResid,
  stat = coursekata::StatReduce,
  position = "identity",
  show.legend = NA,
  show.help = NULL,
  inherit = FALSE,
  environment = parent.frame()
)

gf_squareduce(
  object = NULL,
  gformula = NULL,
  data = NULL,
  ...,
  model,
  aspect = 4/6,
  alpha = 0.1,
  xlab,
  ylab,
  title,
  subtitle,
  caption,
  geom = coursekata::GeomSquareResid,
  stat = coursekata::StatReduce,
  position = "identity",
  show.legend = NA,
  show.help = NULL,
  inherit = FALSE,
  environment = parent.frame()
)

Arguments

Details

Draws squared reduction polygons between the grand mean and the values a fitted model predicts. Each polygon shows one observation's squared reduction; together, their areas represent the model sum of squares. The square is built on the reduction itself and turns with it: a model of the variable the plot puts on x squares the horizontal distance. Its side is scaled to stay square on the page rather than in data units.

Use the same aspect for all three square layers. Across observations, the reduction areas and residual areas sum to the total areas. The equality is between those sums, not between the three squares for any one observation. gf_square_resid(), gf_square_reduce(), and any squared total drawn beside them must use the same aspect for their areas to share a scale.

Value

A ggplot object with squared reduction polygons added.

Examples

set.seed(1)
penguins_20 <- sample(penguins, 20)

# two collections of squares in one sample-level decomposition: squared
# residuals (firebrick) and squared reductions (blue), drawn at one aspect
flipper_model <- lm(body_mass_kg ~ flipper_length_m, data = penguins_20)
gf_point(body_mass_kg ~ flipper_length_m, data = penguins_20) %>%
  gf_model(flipper_model) %>%
  gf_square_resid(flipper_model, color = "firebrick") %>%
  gf_square_reduce(flipper_model, color = "blue")

# and for a two-group model on a jitter plot
gentoo_model <- lm(body_mass_kg ~ gentoo, data = penguins_20)
gf_jitter(body_mass_kg ~ gentoo, data = penguins_20, width = .1) %>%
  gf_model(gentoo_model) %>%
  gf_square_reduce(gentoo_model, color = "blue")

Description

gf_squaresid() is a fully supported alias of gf_square_resid(). The name honors Tyler Haslam, the Utah high school teacher whose efforts shaped the residual and squared-residual visualizations and who requested this function by that name.

Usage

gf_square_resid(
  object = NULL,
  gformula = NULL,
  data = NULL,
  ...,
  model,
  aspect = 4/6,
  alpha = 0.1,
  xlab,
  ylab,
  title,
  subtitle,
  caption,
  geom = coursekata::GeomSquareResid,
  stat = coursekata::StatResid,
  position = "identity",
  show.legend = NA,
  show.help = NULL,
  inherit = FALSE,
  environment = parent.frame()
)

gf_squaresid(
  object = NULL,
  gformula = NULL,
  data = NULL,
  ...,
  model,
  aspect = 4/6,
  alpha = 0.1,
  xlab,
  ylab,
  title,
  subtitle,
  caption,
  geom = coursekata::GeomSquareResid,
  stat = coursekata::StatResid,
  position = "identity",
  show.legend = NA,
  show.help = NULL,
  inherit = FALSE,
  environment = parent.frame()
)

Arguments

Details

Draws squared residual polygons between observed points and the values a fitted model predicts for them, so squared error is an area you can see. The square is built on the residual itself and turns with it: a model of the variable the plot puts on x squares the horizontal distance. Its side is scaled to stay square on the page rather than in data units.

Value

A ggplot object with squared residual polygons added.

Examples

# squared residuals can be drawn on a full data set, but with hundreds of
# points the plot gets hard to read
flipper_model <- lm(body_mass_kg ~ flipper_length_m, data = penguins)
gf_point(body_mass_kg ~ flipper_length_m, data = penguins) %>%
  gf_model(flipper_model) %>%
  gf_square_resid(flipper_model)

# a small sample makes the squared residuals much easier to see
set.seed(1)
penguins_20 <- sample(penguins, 20)

# squared residuals from the empty model (in blue)
empty_model <- lm(body_mass_kg ~ NULL, data = penguins_20)
gf_point(body_mass_kg ~ flipper_length_m, data = penguins_20) %>%
  gf_model(empty_model) %>%
  gf_square_resid(empty_model, color = "blue")

# squared residuals from a two-group model on a jitter plot (in firebrick)
gentoo_model <- lm(body_mass_kg ~ gentoo, data = penguins_20)
gf_jitter(body_mass_kg ~ gentoo, data = penguins_20, width = .1) %>%
  gf_model(gentoo_model) %>%
  gf_square_resid(gentoo_model, color = "firebrick")

# squared residuals from a regression model (in firebrick)
sample_flipper_model <- lm(body_mass_kg ~ flipper_length_m, data = penguins_20)
gf_point(body_mass_kg ~ flipper_length_m, data = penguins_20) %>%
  gf_model(sample_flipper_model) %>%
  gf_square_resid(sample_flipper_model, color = "firebrick")

Description

Usage

gf_square_resid_fun(
  object = NULL,
  gformula = NULL,
  data = NULL,
  ...,
  fun,
  aspect = 4/6,
  alpha = 0.1,
  xlab,
  ylab,
  title,
  subtitle,
  caption,
  geom = coursekata::GeomSquareResid,
  stat = coursekata::StatResid,
  position = "identity",
  show.legend = NA,
  show.help = NULL,
  inherit = FALSE,
  environment = parent.frame()
)

Arguments

Details

Draws squared residual polygons between observed points and the values predicted by a user-supplied function of x. Where gf_square_resid() measures a fitted model, this measures a function you wrote: it is called on the x values the plot draws, and each square is built on the residual from an observation to what the function predicts for it.

Value

A ggplot object with squared residual polygons added.

Examples

set.seed(1)
df <- data.frame(X = 1:10, Y = 2 + 3 * (1:10) + rnorm(10))
my_fun <- function(x) 2 + 3 * x

gf_point(Y ~ X, data = df) %>%
  gf_function(my_fun) %>%
  gf_square_resid_fun(my_fun, color = "red", alpha = 0.3)

Description

Usage

gf_squareplot(
  object = NULL,
  gformula = NULL,
  data = NULL,
  ...,
  binwidth = NULL,
  bins = NULL,
  center = NULL,
  boundary = NULL,
  closed = NULL,
  breaks = NULL,
  bars = "none",
  na.rm = TRUE,
  xlab,
  ylab,
  title,
  subtitle,
  caption,
  geom = coursekata::GeomSquareplot,
  stat = coursekata::StatSquareplot,
  position = "identity",
  show.legend = NA,
  show.help = NULL,
  inherit = TRUE,
  environment = parent.frame()
)

Arguments

Details

Creates histograms where each observation is drawn as its own square, stacked into columns, so a bin's height can be counted as well as read off the axis: n = 47 is 47 squares. Designed for teaching statistical concepts like sampling distributions and hypothesis testing.

Sensible defaults are chosen based on the data:

• For integer-valued data with a small range, the binwidth defaults to 1 so that each integer gets its own column. Everything else about the bins is stat_bin()'s: bins, center, boundary, closed, breaks and pad put a squareplot's columns exactly where a gf_histogram()'s bars would be.

• A factor keeps its levels, so a level nobody landed in still holds its place on the axis. Knowing a value never occurred is the point.

• A discrete x – a factor, character or logical vector – is counted, one column per level, positioned the way gf_bar() positions its bars. The binning arguments (bins, binwidth, center, boundary, closed, breaks, pad) belong to a continuous x; supplying one alongside a discrete x warns rather than changing the plot.

• The white separator between two squares is capped at a quarter of a square's smaller side, so as a bin fills and its squares shrink the separator thins with them and the squares stay countable.

• The y axis is a count, so its breaks are whole numbers.

• Counts must be finite, nonnegative whole numbers. Integer frequency weights are supported; fractional weighted counts need a histogram or bar plot instead.

Fill and separator colour inherit the plot's mappings. With no mapping or fixed value, squares use teal fill and white separators. Set fill or color explicitly to override an inherited mapping. Plot-level after_scale() linewidth mappings use automatic border fitting. Supply the mapping on the squareplot layer itself to keep its width fixed.

The bins are a histogram's bins: binwidth, bins, center, boundary, closed and breaks mean what they mean on ggformula::gf_histogram(), and the same arguments give the same bin edges and the same counts, so squares laid over bars land inside them.

Everything that is not the squares is a layer or a scale: ⁠%>% show_mean()⁠, ⁠%>% show_dgp()⁠, ⁠%>% gf_lims(x = )⁠, ⁠%>% gf_refine(ggplot2::expand_limits(y = ))⁠. Each of these was an argument here once, and passing the old name is refused with the replacement named, so a call written against the old signature says what to write rather than drawing a plot with the mark missing.

Value

A ggplot object.

See Also

show_mean() and show_dgp() annotate a distribution. The sampling distributions guide shows this plot in the context of a full shuffle-and-estimate workflow: https://coursekata.github.io/coursekata-r/articles/sampling-distributions.html

Examples

# each observation is a countable square
gf_squareplot(~Thumb, data = Fingers)

# `bars` controls the display: "none" (default), "outline", or "solid"
gf_squareplot(~Thumb, data = Fingers, bars = "outline")

# the bins are a histogram's bins, so squares laid over bars land inside them --
# name the grid on both layers, because a layer never reads its neighbor's
gf_histogram(~Thumb, data = Fingers, bins = 8) %>% gf_squareplot(bins = 8)

# customize fill color, binwidth, and axis limits
gf_squareplot(~Thumb, data = Fingers, fill = "coral", binwidth = 5) %>%
  gf_lims(x = c(30, 90))

# integer data with a small range gets one column per integer
int_data <- data.frame(rolls = sample(1:6, 30, replace = TRUE))
gf_squareplot(~rolls, data = int_data)

# the plot is a real ggformula layer, so it facets and takes mapped aesthetics
gf_squareplot(~ Thumb | Sex, data = Fingers)
gf_squareplot(~Thumb, data = Fingers, fill = ~Sex)

# with 2000 observations the squares shrink, and their separators thin to fit
set.seed(24)
large_data <- data.frame(x = rnorm(2000, mean = 50, sd = 10))
gf_squareplot(~x, data = large_data)

# show a dashed line at the sample mean
gf_squareplot(~Thumb, data = Fingers) %>% show_mean()

# frame a sampling distribution with its data generating process: with only
# 10 shuffles, the mean of the distribution can land far from the null.
# Position-scale limits and show_dgp() compose in either order; neither
# changes the count range.
shuffled_b1 <- function(n) {
  data.frame(b1 = replicate(n, {
    shuffled_tip <- base::sample(TipExperiment$Tip)
    b1(lm(shuffled_tip ~ Condition, data = TipExperiment))
  }))
}

set.seed(42)
gf_squareplot(~b1, data = shuffled_b1(10), binwidth = 2) %>%
  gf_lims(x = c(-30, 30)) %>%
  gf_refine(ggplot2::expand_limits(y = 10)) %>%
  show_mean() %>%
  show_dgp()

# a factor keeps every level, including the ones nothing landed in
ratings <- data.frame(rating = factor(
  base::sample(1:5, 20, replace = TRUE, prob = c(1, 2, 4, 2, 1)),
  levels = 1:5
))
gf_squareplot(~rating, data = ratings)

Description

guide_cutoff() draws one truthful marker rail and, when requested, one measured label lane per helper call. Two-sided labels dodge along that lane while their leaders remain tied to the truthful cutoff anchors. The guide receives already-computed cutoff values; it never inspects the distribution or recomputes a statistic.

Usage

guide_cutoff(
  value,
  label = NULL,
  side = NULL,
  call_id = 1L,
  colour = ggplot2::waiver(),
  shape = 24,
  size = 4,
  linewidth = 0.5,
  title = NULL,
  theme = NULL,
  order = 0,
  position = ggplot2::waiver()
)

GuideCutoff

Arguments

Format

GuideCutoff is a ggplot2::Guide object.

Value

A ggplot2 guide.

Description

guide_dgp() places a truthful null anchor in measured guide space. The population role supplies the population heading, equation, axis line, null label, and triangle. The estimate role supplies the estimate heading, equation, and null label. It never computes a mean or trains a scale.

Usage

guide_dgp(
  value = 0,
  role = c("population", "estimate"),
  label = ggplot2::waiver(),
  equation = ggplot2::waiver(),
  colour = ggplot2::waiver(),
  shape = ggplot2::waiver(),
  size = ggplot2::waiver(),
  linewidth = ggplot2::waiver(),
  title = ggplot2::waiver(),
  theme = NULL,
  order = 0,
  position = ggplot2::waiver()
)

GuideDgp

Arguments

Format

GuideDgp is a ggplot2::Guide object.

Value

A ggplot2 guide.

Description

Given a distribution, find which values lie in the upper, lower, or middle proportion of the distribution. Useful when you want to do something like shade in the middle 95% of a plot. This is a greedy operation, meaning that if the cutoff point is between two whole numbers the specified region will suck up the extra space. For example, the requesting the upper 30% of the ⁠[1 2 3 4]⁠ will return ⁠[FALSE FALSE TRUE TRUE]⁠ because the 30% was greedy.

outer() marks values in both outer tails of a distribution. It is the complement of middle(): outer(x, prop) is equivalent to tails(x, 1 - prop).

Usage

middle(x, prop = 0.95, greedy = TRUE)

outer(x, prop)

tails(x, prop = 0.95, greedy = TRUE)

lower(x, prop = 0.025, greedy = TRUE)

upper(x, prop = 0.025, greedy = TRUE)

Arguments

Details

Note that NA values are ignored when sizing the region, and come back as NA.

Value

A logical vector indicating which values are in the specified region.

See Also

The sampling distributions guide walks through building these distributions with do() and shuffle(), and shows the bootstrap variant: https://coursekata.github.io/coursekata-r/articles/sampling-distributions.html

Examples

# each function returns a logical vector marking the values in its region
upper(1:10, .1)
lower(1:10, .2)
middle(1:10, .5)
tails(1:10, .5)

# they are most often used as the fill aesthetic of a histogram of a
# sampling distribution -- here, b1s estimated from shuffled (null) data
set.seed(42)
shuffled <- data.frame(b1 = replicate(200, {
  shuffled_tip <- base::sample(TipExperiment$Tip)
  b1(lm(shuffled_tip ~ Condition, data = TipExperiment))
}))

# color the middle 95%: the b1 values we would expect to see often
# if the empty model were true
gf_histogram(~b1, data = shuffled, binwidth = 1, fill = ~ middle(b1, .95))

# tails() marks the same cutoffs with the opposite coloring: the values
# outside the middle 95% are the 5% most extreme
gf_histogram(~b1, data = shuffled, binwidth = 1, fill = ~ tails(b1, .95))

# outer() marks the same region as tails() but takes the tail proportion
# directly: the outer 5%
gf_histogram(~b1, data = shuffled, binwidth = 1, fill = ~ outer(b1, .05))

# upper() and lower() are for directional hypotheses: all 5% goes in one tail
gf_histogram(~b1, data = shuffled, binwidth = 1, fill = ~ upper(b1, .05))
gf_histogram(~b1, data = shuffled, binwidth = 1, fill = ~ lower(b1, .05))

A modified form of the palmerpenguins::penguins data set.

Description

The modifications are to select only a subset of the variables, and convert some of the units.

Usage

penguins

Format

A data frame with 333 observations on the following 7 variables:

species

The species of penguin, coded as "Adelie", "Chinstrap", or "Gentoo".

gentoo

Whether the penguin is a Gentoo penguin (1) or not (0).

body_mass_kg

The mass of the penguin's body, in kilograms.

flipper_length_m

The length of the penguin's flipper, in m.

bill_length_cm

The length of the penguin's bill, in cm.

female

Whether the penguin is female (1) or not (0).

island

The island where the penguin was observed, coded as "Biscoe", "Dream", or "Torgersen".

Description

See coursekata_palette() for more information.

Usage

scale_discrete_coursekata(...)

Arguments

Value

A discrete color scale.

See Also

coursekata_palette

Examples

gf_point(Thumb ~ Height, data = Fingers, color = ~RaceEthnic) +
  scale_discrete_coursekata()

Description

Usage

show_cutoffs(
  object = NULL,
  part,
  color = "#1e3a8a",
  size = 4,
  show_labels = FALSE
)

Arguments

Details

Adds empirical quantile cutoffs for a distribution part – middle(), tails(), upper(), lower(), or outer(). Short dashed stems end in downward triangles that touch the numeric axis. Optional callouts are drawn in the data panel. Repeated labelled calls are measured and placed together. Their boxes stay separate when the panel has room, and their routed leaders remain tied to the exact cutoffs.

In very small panels, leaders use a simple elbow connection if there is no room for an outward route. Labels keep their text and size, so they may overlap or be clipped. Enlarge the figure, use fewer facet columns, or set show_labels = FALSE when labels do not fit.

By default the part is read off the plot's fill aesthetic, e.g. fill = ~middle(Thumb, .95). Passing part overrides that reading: it marks whatever part is named there instead, and the fill (if any) is ignored – marking the 99% cutoffs on a plot shaded for the 95% is a deliberate, lossless override, not a mismatch.

show_cutoffs() refuses a plot whose first layer does not draw a distribution (a scatterplot, for instance) and, when part is given explicitly, a part that names a variable other than the one the plot puts on x.

Value

A ggplot object with cutoff markers and optional labels.

See Also

stat_cutoff() computes the same rule per panel; geom_cutoff() is the lower-level drawing component. guide_cutoff() can place cutoff anchors in a position guide instead.

Examples

gf_histogram(~Thumb, data = Fingers, binwidth = 5, fill = ~middle(Thumb, .95)) %>%
  show_cutoffs(show_labels = TRUE)

# an explicit part overrides the fill instead of requiring it to match
gf_histogram(~Thumb, data = Fingers, binwidth = 5, fill = ~middle(Thumb, .95)) %>%
  show_cutoffs(middle(Thumb, .99))

Description

Usage

show_dgp(object = NULL, color = "#003d70", null_color = "#E60000", size = 4)

Arguments

Details

Frames a distribution of estimates with the process that generated them: the population model on a top axis labeled "Population Parameter (DGP)", the sample estimate below the plot, and a marker at the null hypothesis ($\beta_1 = 0$) on both – drawn only when zero is on the axis.

The population and estimate frames are position guides, outside the data panel. They therefore do not change the count range and remain compatible with fixed, zoomed, transformed, and free count axes. The ordinary numeric x guide stays in place as the estimate scale. Zero is omitted when it is not a finite visible value on that scale; it is never moved to a boundary. An x position scale can be added before or after show_dgp(); the final scale supplies the limits and breaks for both guides. A later ggplot2::guides() call can replace either guide in the usual ggplot2 way. The plot's x label is retained when a later guide replaces the estimate frame. Existing plot theme settings, including inherited axis.text, axis.title, and axis.line settings, take precedence over the frame's default axis styling. Later themes follow ggplot2's usual inheritance: change axis.text.x, axis.title.x, or axis.line.x/axis.line.y to override the frame's axis-specific settings.

Value

The plot, with population and estimate guides added.

See Also

The sampling distributions guide draws this figure inside a full shuffle-and-estimate workflow: https://coursekata.github.io/coursekata-r/articles/sampling-distributions.html

Examples

# with only ten shuffles the mean of the distribution can land well away
# from the null hypothesis marked on the top axis
set.seed(42)
shuffled <- data.frame(b1 = replicate(10, {
  b1(lm(base::sample(TipExperiment$Tip) ~ Condition, data = TipExperiment))
}))

# expand_limits() sets the count axis so two runs can be compared side by
# side; show_dgp() does not alter that range
gf_histogram(~b1, data = shuffled, binwidth = 2) %>%
  gf_refine(ggplot2::expand_limits(y = 10)) %>%
  show_mean() %>%
  show_dgp()

Description

Usage

show_mean(
  object = NULL,
  color = "#E60000",
  linetype = "longdash",
  linewidth = 0.7
)

Arguments

Details

Draws a vertical line at the mean of the variable a distribution is built from, spanning the count axis. Works on any plot of one distribution – gf_histogram(), gf_dotplot(), gf_squareplot(). A faceted plot gets one line per panel, at that panel's own mean, because a facet is a region with its own subset of the data. Each line spans the count axis of the panel it is drawn in, so scales = "free_y" is supported and no panel is stretched to hold another panel's line.

A plot with a variable on each axis is refused. The middle of a two-variable plot is a model, and gf_model() draws models – an hline at the mean of the outcome for the empty model. Averaging the x variable of a scatterplot would draw a line nobody asked for.

Value

The plot, with a tagged mean line added.

See Also

show_dgp() frames a sampling distribution with the process that generated it; gf_model() draws the mean of a two-variable plot's outcome.

Examples

gf_histogram(~Thumb, data = Fingers, binwidth = 5) %>% show_mean()

# a facet is a region with its own subset, so each panel gets its own mean
gf_histogram(~Thumb | Sex, data = Fingers, binwidth = 5) %>% show_mean()

Description

These data are simulated to be similar to the Ames housing data, but with far fewer variables and much smaller effect sizes.

Usage

Smallville

Format

A data frame with 32 observations on the following 4 variables:

PriceK

Price the home sold for (in thousands of dollars)

Neighborhood

The neighborhood the home is in (Eastside, Downtown)

HomeSizeK

The size of the home (in thousands of square feet)

HasFireplace

Whether the home has a fireplace (0 = no, 1 = yes)

Description

Split data into train and test sets.

Usage

split_data(data, prop = 0.7)

Arguments

Value

A list with two data frames, train and test.

Examples

set.seed(123)
parts <- split_data(Fingers, prop = 0.8)
nrow(parts$train)
nrow(parts$test)

Description

StatCutoff calls the same cutoff_plan() used by show_cutoffs(), but the two interfaces deliberately operate at different scopes. ggplot2 supplies a stat with one panel's rows at a time, so this stat computes panel-local cutoffs. show_cutoffs() plans once from the whole resolved distribution and repeats those global marks across facets.

Usage

StatCutoff

stat_cutoff(
  mapping = NULL,
  data = NULL,
  geom = "cutoff",
  position = "identity",
  ...,
  part = "middle",
  prop = 0.95,
  greedy = TRUE,
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE
)

Arguments

Format

StatCutoff is a ggplot2::Stat object.

Details

The stat accepts "middle", "tails", "upper", "lower", and "outer". It selects cutoffs in the data's original space: transformed position values are inverted before cutoff_plan() runs, and the selected cutoffs are transformed forward again for ggplot2. This preserves the requested distribution part on increasing, decreasing, and transformed scales.

stat_cutoff() is the conventional layer constructor. Its default geom is "cutoff"; use geom = "vline" when a full-height reference line is wanted instead. Since the stat computes once per panel, styling aesthetics from the source data cannot be mapped; set them to one value, or facet the plot to compute one set of cutoffs per group.

Value

stat_cutoff() returns a ggplot2 layer.

See Also

show_cutoffs(), which marks whole-distribution cutoffs.

Description

StatDistMean receives the rows that ggplot2 has assigned to each panel and emits their mean as an xintercept. Because ggplot2 transforms position scales and applies hard-limit out-of-bounds handling before a stat runs, the result follows that lifecycle: a transformed scale changes the values being averaged, hard scale limits can remove values, and coordinate zoom does not change the result.

Usage

StatDistMean

stat_dist_mean(
  mapping = NULL,
  data = NULL,
  geom = "vline",
  position = "identity",
  ...,
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE
)

Arguments

Format

StatDistMean is a ggplot2::Stat object.

Details

stat_dist_mean() is the conventional layer constructor. It computes from x and defaults to a vertical line; show_mean() is the teaching-oriented helper that resolves a distribution plot's mapping and styles this layer. Since the stat computes once per panel, styling aesthetics from the source data cannot be mapped; set them to one value, or facet the plot to compute one mean per group.

Value

stat_dist_mean() returns a ggplot2 layer.

See Also

show_mean() and gf_model().

Description

StatModel passes through rows prepared from a supplied model. With no supplied model, it fits the model implied by the mapped positions: a linear model for a continuous predictor, group means for a categorical predictor, or an intercept when only the outcome is mapped.

Usage

StatModel

Format

StatModel is a ggplot2::Stat object.

Description

A reduction uses the same endpoint representation as a residual, but its starting point is the grand mean rather than the observation. The separate stat class keeps that statistical meaning visible to ggplot2 users and to code inspecting the layer. GeomResid and GeomSquareResid can draw either stat.

Usage

StatReduce

Format

A ggplot2::Stat object.

See Also

stat_reduce(), geom_reduce(), and geom_square_reduce().

Description

xend/yend is a positional aesthetic on purpose: that is what gets the prediction transformed by its scale and trained into the panel's range. Exactly one of the two arrives, on the axis the plot put the model's outcome on.

Usage

StatResid

Format

A ggplot2::Stat object.

Details

compute_layer checks the required aesthetics, then returns every row unchanged. An observation with an NA on it (a predictor the model dropped) survives the stat instead of being removed before the position runs. The residual's jitter is a function of the seed and the rows it is handed, exactly as the point layer's is, so losing a row here would hand it a different sequence of draws and the segment would land away from its point.

See Also

stat_resid(), geom_resid(), and geom_square_resid().

Description

StatSdRuler reduces each panel to the segment for one standard deviation, anchored at the outcome's mean. With x and y aesthetics, the outcome is y; the ruler is vertical and where places it along x. With x alone, the outcome is x and the ruler runs along the baseline from the mean to one standard deviation above it.

Usage

StatSdRuler

stat_sd_ruler(
  mapping = NULL,
  data = NULL,
  geom = "segment",
  position = "identity",
  ...,
  where = "middle",
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE
)

Arguments

Format

A ggplot2::Stat object.

Details

stat_sd_ruler() is the conventional layer constructor. It uses the same stat as gf_sd_ruler(), so both interfaces measure transformed values and compute one ruler from each panel's rows. Groups within a panel do not get separate rulers. Styling aesthetics from the source data cannot be mapped; set them to one value, or facet the plot to draw one ruler per group.

Value

stat_sd_ruler() returns a ggplot2 layer.

See Also

gf_sd_ruler() provides the ggformula interface.

Examples

ggplot2::ggplot(Fingers, ggplot2::aes(Height, Thumb)) +
  ggplot2::geom_point() +
  stat_sd_ruler(where = "mean", colour = "red", linewidth = 1)

ggplot2::ggplot(Fingers, ggplot2::aes(Thumb)) +
  ggplot2::geom_histogram(bins = 30) +
  stat_sd_ruler(colour = "red", linewidth = 1)

Description

StatSquareplot is ggplot2::StatBin: binwidth, bins, center, boundary, closed, breaks and pad all mean exactly what they mean on a histogram, because they are the histogram's own code. GeomSquareplot re-expands each bin into one rectangle per observation. Pair the two in a ggplot2::layer() to put countable squares in a plot you are assembling yourself.

Usage

StatSquareplot

Format

A ggplot2::Stat object.

Details

The one difference from a plain histogram is the binwidth chosen when the call names no grid at all: integer-valued data over a small range gets one bin per integer, rather than stat_bin's bins = 30 default, because thirty slivers is the wrong advice for a plot whose point is countability.

See Also

gf_squareplot(), which pairs this stat and geom for you.

Description

Students at a university taking an introductory statistics course were asked to complete this survey as part of their homework.

Usage

Survey

Format

A data frame with 211 observations on the following 1 variable:

Any1_20

The random number between 1 and 20 that a student thought of.

Description

Data about tips collected from an experiment with 44 tables at a restaurant.

Usage

Tables

Format

A data frame with 44 observations on the following 2 variables.

TableID

A number assigned to each table.

Tip

How much the tip was.

A simple theme built on top of ggplot2::theme_bw

Description

The coursekata package automatically loads this theme when the package is loaded. This is in addition to a number of other plot tweaks and option settings. To just restore the theme to the default, you can run set_theme(theme_grey). If you want to restore all plot related settings and/or prevent them when loading the package, see coursekata_unload_theme.

Usage

theme_coursekata()

Value

A gg theme object

Examples

gf_boxplot(Thumb ~ RaceEthnic, data = Fingers, fill = ~RaceEthnic)

Description

These are simulated data that are similar to the TipExperiment data. Hypothetical tables were randomly assigned to receive checks that either included or did not include a drawing of a smiley face, either from a male or a female server.

Usage

tip_exp

Format

A data frame with 44 observations on the following 3 variables.

gender

Whether the server was female or male

condition

Whether the check had a ⁠smiley face⁠ or not (control)

tip_percent

The size of the tip as a percentage of the price of the meal

Description

Tables were randomly assigned to receive checks that either included or did not include a drawing of a smiley face. Data was collected from 44 tables in an effort to examine whether the added smiley face would cause more generous tipping.

Usage

TipExperiment

Format

A data frame with 44 observations on the following 3 variables.

TableID

A number assigned to each table.

Tip

How much the tip was.

Condition

Which experimental condition the table was randomly assigned to.

Check

(Simulated) The amount of money the table paid for their meal.

FoodQuality

(Simulated) The perceived quality of the food.

Description

These data have been updated with some historical height data (from Our World in Data), drinking data (collected by the World Health Organization and published with the FiveThirtyEight alcohol-consumption data), population and land characteristics, and vaccination data (from March 2023).

Usage

World

Format

A data frame with 130 observations on the following 14 variables:

Country

Name of country

Region

One of 5 UN defined regions: Africa, Americas, Asia, Europe, Oceania

Code

Three-letter country codes defined by the International Organization for Standardization (ISO) to represent countries in a way that avoids errors since a country’s name changes depending on the language being used.

LifeExpectancy

Average life expectancy (in years)

GirlsH1900

The average of 18-year-old girls heights in 1900 (in cm)

GirlsH1980

The average of 18-year-old girls heights in 1980 (in cm)

Happiness

Score on a 0-10 scale for average level of happiness (10 being happiest)

GDPperCapita

Gross Domestic Product (per capita)

FertRate

The average number of children that will be born to a woman over her lifetime

PeopleVacc

Total number of people vaccinated in the country

PeopleVacc_per100

Total number of people vaccinated in the country (in percent)

Population2010

Population (in millions) in 2010

Population2020

Population (in millions) in 2020

WineServ

Average wine consumption per capita for those age 15 and over per week (collected by WHO)