Package 'spinifex'

Description

Internal function. To be applied to aes_args replicates elements to the length of the data and bind as a column.

Usage

.bind_elements2df(list, df)

Arguments

See Also

Other Internal utility: .init4proto, .lapply_rep_len()

Examples

## This function is not meant for external use

Description

Internal expression. Creates local .objects to be commonly consumed by spinifex proto_* functions.

Usage

.init4proto

Format

An object of class expression of length 1.

See Also

Other Internal utility: .bind_elements2df(), .lapply_rep_len()

Examples

## This expression. is not meant for external use.

Description

Internal function. To be applied to aes_args and identity_args, replicates vectors of length data to length of data*frames for animation.

Usage

.lapply_rep_len(list, to_length, expected_length)

Arguments

See Also

Other Internal utility: .bind_elements2df(), .init4proto

Examples

## This function is not meant for external use

Description

Animates the ggplot return of ggtour() and added ⁠proto_*()⁠ functions as a .gif without interaction, through use of {gganimate}.

Usage

animate_gganimate(
  ggtour,
  fps = 8,
  rewind = FALSE,
  start_pause = 1,
  end_pause = 1,
  ...
)

Arguments

See Also

gganimate::animate

Other ggtour animator: animate_plotly(), filmstrip()

Examples

library(spinifex)
dat     <- scale_sd(penguins_na.rm[, 1:4])
clas    <- penguins_na.rm$species
bas     <- basis_pca(dat)
mv      <- manip_var_of(bas)
mt_path <- manual_tour(bas, manip_var = mv)

ggt <- ggtour(mt_path, dat, angle = .3) +
  proto_default(aes_args = list(color = clas, shape = clas),
                identity_args = list(size = 1.5, alpha = .7))
## Not run: 
## Default .gif rendering
animate_gganimate(ggt)

if(FALSE){ ## Don't accidentally save file
  ## Option arguments, rendering to default .gif
  anim <- animate_gganimate(
    ggt, fps = 10, rewind = TRUE,
    start_pause = 1, end_pause = 2,
    height = 10, width = 15, units = "cm", ## "px", "in", "cm", or "mm."
    res = 200 ## resolution, pixels per dimension unit I think
  )
  ## Save rendered animation
  gganimate::anim_save("my_tour.gif",
                       animation = anim,
                       path = "./figures")
  
  ## Alternative renderer saving directly to .mp4
  animate_gganimate(ggt, fps = 5,
    height = 4, width = 6, units = "in", ## "px", "in", "cm", or "mm."
    res = 200, ## resolution, pixels per dimension unit I think
    renderer = gganimate::av_renderer("./my_tour.mp4"))
}

## End(Not run)

Description

Animates the static ggtour() and added ⁠proto_*()⁠ functions as a {plotly} animation, an .html widget with slider and hover tooltip showing the row number.

Usage

animate_plotly(ggtour, fps = 8, ...)

Arguments

See Also

Other ggtour animator: animate_gganimate(), filmstrip()

Examples

library(spinifex)
dat     <- scale_sd(penguins_na.rm[, 1:4])
clas    <- penguins_na.rm$species
bas     <- basis_pca(dat)
mv      <- manip_var_of(bas)
mt_path <- manual_tour(bas, manip_var = mv)

ggt <- ggtour(mt_path, dat, angle = .3) +
  proto_origin() +
  proto_basis() +
  proto_point(aes_args = list(color = clas, shape = clas),
              identity_args = list(size = 1.5, alpha = .7))

animate_plotly(ggt, width = 700, height = 450) ## pixels only, no resolution argument

## Example saving to a .html widget, may require additional setup.
if(FALSE){
  anim <- animate_plotly(ggt, fps = 10,
                         width = 700, height = 450) ## in pixels
  
  htmlwidgets::saveWidget(widget = anim,
                          file = "./figures/my_tour.html",
                          selfcontained = TRUE)}

Description

Adds/overwrites the y of the projected data. Usefully for 1D projections and appending information related to, but independent from the projection; model predictions or residuals for instance. Wants to be called early so that the following proto calls adopt the changes.

Usage

append_fixed_y(fixed_y)

Arguments

See Also

Other ggtour proto functions: facet_wrap_tour(), ggtour(), proto_basis(), proto_default(), proto_density(), proto_density2d(), proto_hex(), proto_highlight(), proto_hline0(), proto_origin(), proto_point(), proto_text()

Examples

library(spinifex)
dat     <- scale_sd(penguins_na.rm[, 1:4])
clas    <- penguins_na.rm$species
bas     <- basis_pca(dat)
mv      <- manip_var_of(bas)
mt_path <- manual_tour(bas, manip_var = mv)

# Fixed y height with related information, independent of a 1D tour 
# _eg_ predictions or residuals.
message("don't forget to scale your fixed_y.")
dummy_y <- scale_sd(as.integer(clas) + rnorm(nrow(dat), 0, .5))
gt_path <- save_history(dat, grand_tour(d = 1), max_bases = 5)

message("append_fixed_y wants to be called early so other proto's adopt the fixed_y.")
ggt <- ggtour(gt_path, dat, angle = .3) +
  append_fixed_y(fixed_y = dummy_y) + ## insert/overwrites vertical values.
  proto_point(list(fill = clas, color = clas)) +
  proto_basis1d() +
  proto_origin()

animate_plotly(ggt)

Description

Internal function, many end users will not need this. Takes the result of manual_tour() or tourr::save_history(). Restructures the array of interpolated bases into a long data frame for use in ggplots.

Usage

array2df(
  basis_array,
  data = NULL,
  basis_label = NULL,
  data_label = rownames(data),
  do_center_frame = TRUE
)

Arguments

Examples

## !!This function is not meant for external use!!
dat  <- scale_sd(wine[, 2:6])
clas <- wine$Type
bas  <- basis_pca(dat)
mv   <- manip_var_of(bas)

## Radial tour array to long df, as used in play_manual_tour()
mt_array <- manual_tour(basis = bas, manip_var = mv)
ls_df_frames <- array2df(basis_array = mt_array, data = dat,
                         basis_label = paste0("MyLabs", 1:nrow(bas)))
str(ls_df_frames)

## tourr::save_history tour array to long df, as used in play_tour_path()
gt_array <- tourr::save_history(data = dat, max_bases = 10)
ls_df_frames2 <- array2df(basis_array = gt_array, data = dat)
str(ls_df_frames2)

Description

Performs simulated annealing on the index function, solving for it's local extrema. Returns only the last identified basis of the optimization. A truncated, muted extension of tourr::save_history(guided_tour())).

Usage

basis_guided(data, index_f = tourr::holes(), d = 2, ...)

Arguments

Value

Numeric matrix of the last basis of a guided tour.

See Also

tourr::guided_tour for annealing arguments.

Other basis producing functions: basis_half_circle(), basis_odp(), basis_olda(), basis_onpp(), basis_pca()

Examples

dat <- scale_sd(wine[, 2:6])
basis_guided(data = dat, index_f = tourr::holes())

basis_guided(data = dat, index_f = tourr::cmass(),
             alpha = .4, cooling = .9, max.tries = 10, n_sample = 4)

Description

Orthonormalizes uniform variable contributions on a unit circle. This serves as a NULL basis, one that is variable agnostic while spacing the variables to have minimize variable dependence.

Usage

basis_half_circle(data)

Arguments

See Also

Other basis producing functions: basis_guided(), basis_odp(), basis_olda(), basis_onpp(), basis_pca()

Examples

dat <- scale_sd(wine[, 2:6])
bas <- basis_half_circle(dat)

Description

Orthogonal Discriminant Projection (ODP) is a linear dimension reduction method with class supervision. It maximizes weighted difference between local and non-local scatter while local information is also preserved by constructing a neighborhood graph.

Usage

basis_odp(data, class, d = 2, type = c("proportion", 0.1), ...)

Arguments

References

Li B, Wang C, Huang D (2009). "Supervised feature extraction based on orthogonal discriminant projection." Neurocomputing, 73(1-3), 191-196.

See Also

Rdimtools::do.odp for locality preservation arguments.

Rdimtools::aux.graphnbd for details on type.

Other basis producing functions: basis_guided(), basis_half_circle(), basis_olda(), basis_onpp(), basis_pca()

Examples

dat  <- scale_sd(wine[, 2:6])
clas <- wine$Type
basis_odp(data = dat, class = clas)

Description

Orthogonal LDA (OLDA) is an extension of classical LDA where the discriminant vectors are orthogonal to each other.

Usage

basis_olda(data, class, d = 2)

Arguments

Value

A numeric matrix, an orthogonal basis that best distinguishes the group means of class.

References

Ye J (2005). "Characterization of a Family of Algorithms for Generalized Discriminant Analysis on Undersampled Problems." J. Mach. Learn. Res., 6, 483-502. ISSN 1532-4435.

See Also

Rdimtools::do.olda

Other basis producing functions: basis_guided(), basis_half_circle(), basis_odp(), basis_onpp(), basis_pca()

Examples

dat  <- scale_sd(wine[, 2:6])
clas <- wine$Type
basis_olda(data = dat, class = clas)

Description

Orthogonal Neighborhood Preserving Projection (ONPP) is an unsupervised linear dimension reduction method. It constructs a weighted data graph from LLE method. Also, it develops LPP method by preserving the structure of local neighborhoods. For the more details on type see Rdimtools::aux.graphnbd().

Usage

basis_onpp(data, d = 2, type = c("knn", sqrt(nrow(data))))

Arguments

Value

Orthogonal matrix basis that distinguishes the levels of class based on local and non-local variation as weighted against the neighborhood graph.

References

He X (2005). Locality Preserving Projections. PhD Thesis, University of Chicago, Chicago, IL, USA.

See Also

Rdimtools::do.onpp

Rdimtools::aux.graphnbd for details on type.

Other basis producing functions: basis_guided(), basis_half_circle(), basis_odp(), basis_olda(), basis_pca()

Examples

dat  <- scale_sd(wine[, 2:6])
basis_onpp(data = dat)

Description

The orthogonal linear components of the variables in the next largest direction of variance.

Usage

basis_pca(data, d = 2)

Arguments

See Also

Rdimtools::do.pca

Other basis producing functions: basis_guided(), basis_half_circle(), basis_odp(), basis_olda(), basis_onpp()

Examples

dat <- scale_sd(wine[, 2:6])
basis_pca(data = dat)

Description

The objective is to identify each of a number of benign or malignant classes. Samples arrive periodically as Dr. Wolberg reports his clinical cases. The database therefore reflects this chronological grouping of the data. This grouping information appears immediately below, having been removed from the data itself. Each variable except for the first was converted into 11 primitive numerical attributes with values ranging from 0 through 10. Rows with missing attribute values and duplicate rows removed.

Usage

BreastCancer_na.rm

Format

A data frame with 675 observations of 8 numeric variables and target factor Class.

• Id, Sample code number

• Cl.thickness, Clump thickness

• Cell.size, Uniformity of cell size

• Cell.shape, Uniformity of cell shape

• Marg.adhesion, Marginal adhesion

• Epith.c.size, Single Epthelial cell size

• Bare.nuclei, Bare nuclei

• Bl.cromatin, Bland chromatin

• Normal.nucleoli, Normal Nucleoli

• Mitoses, Mitoses

• Class, Class of cancer, either "benign" or "malignant"

Details

This is a cleaned subset of mlbench's BreastCancer. See help(BreastCancer, package = "mlbench") for the original.

Replicating this dataset:

require("mlbench")
data(BreastCancer)

raw <- BreastCancer
## rownumber index of 8 duplicate 16 incomplete rows
idx <- !duplicated(raw) & complete.cases(raw) 
d <- raw[idx, 3:10]
d <- apply(d, 2L, as.integer)
d <- data.frame(d, Class = as.factor(raw$Class[idx]))
BreastCancer_na.rm <- d
## save(BreastCancer_na.rm, file = "./data/BreastCancer_na.rm.rda")

Source

J.W. Smith., el al. 1988. Using the ADAP learning algorithm to forecast the onset of diabetes mellitus. In Proceedings of the Symposium on Computer Applications and Medical Care (pp. 261–265). IEEE Computer Society Press.

mlbench, R package. F. Leisch & E. Dimitriadou, 2021. mlbench: Machine Learning Benchmark Problems https://CRAN.R-project.org/package=mlbench

Examples

library(spinifex)
str(BreastCancer_na.rm)
dat  <- scale_sd(BreastCancer_na.rm[, 1:8])
clas <- BreastCancer_na.rm$Class

bas <- basis_pca(dat)
mv  <- manip_var_of(bas)
mt  <- manual_tour(bas, mv)

ggt <- ggtour(mt, dat, angle = .2) +
  proto_default(aes_args = list(color = clas, shape = clas))

animate_plotly(ggt)

Description

Typically called by manual_tour(). Creates a (p, d) orthonormal matrix, the manipulation space from the given basis right concatenated with a zero vector, with manip_var set to 1.

Usage

create_manip_space(basis, manip_var = manip_var_of(basis))

Arguments

Value

A (p, d + 1) orthonormal matrix, the manipulation space to manipulate the projection in.

See Also

Other manual tour adjacent functions: interpolate_manual_tour(), manip_var_of(), manual_tour(), rotate_manip_space()

Examples

library(spinifex)
dat <- scale_sd(wine[, 2:6])
bas <- basis_pca(dat)
mv  <- manip_var_of(bas)
create_manip_space(basis = bas, manip_var = mv)

## d = 1 case 
bas1d <- basis_pca(dat, d = 1)
mv    <- manip_var_of(bas1d)
create_manip_space(bas1d, mv)

Description

Send a message if the 4th chunk of the package version is 9000.

Usage

devMessage(text)

Arguments

Description

Additively draws a basis on a static ggplot. Not a geom or proto. Expects

Usage

draw_basis(
  basis,
  map_to = data.frame(x = c(0, 1), y = c(0, 1)),
  position = c("left", "center", "right", "bottomleft", "topright", "off"),
  manip_col = "blue",
  line_size = 0.6,
  text_size = 4,
  basis_label = abbreviate(gsub("[^[:alnum:]=]", "", rownames(basis), 3L))
)

Arguments

Examples

library(spinifex)
library(ggplot2)
dat  <- scale_sd(penguins_na.rm[, 1:4])
clas <- penguins_na.rm$species
bas  <- basis_pca(dat)
proj <- as.data.frame(dat %*% bas)

ggplot() +
  geom_point(aes(PC1, PC2), proj) +
  draw_basis(bas, proj, "left") +
  coord_fixed()
  
## Aesthetics and basis on specific facet levels
proj <- cbind(proj, clas = penguins_na.rm$species)
bas <- cbind(as.data.frame(bas), clas = levels(clas)[2])
ggplot() +
  facet_wrap(vars(clas)) +
  geom_point(aes(PC1, PC2, color = clas, shape = clas), proj) +
  draw_basis(bas, proj, "left") +
  theme_spinifex()
## To repeat basis in all facet levels don't cbind a facet variable.

Description

Create and wrap a 1d ribbon of panels in 2d. Because of the side effects of ggtour and facet_wrap_tour this wants to be applied after ggtour and before any ⁠proto_*⁠ functions. plotly may not display well with with faceting.

Usage

facet_wrap_tour(facet_var, nrow = NULL, ncol = NULL, dir = "h")

Arguments

See Also

Other ggtour proto functions: append_fixed_y(), ggtour(), proto_basis(), proto_default(), proto_density(), proto_density2d(), proto_hex(), proto_highlight(), proto_hline0(), proto_origin(), proto_point(), proto_text()

Examples

library(spinifex)
dat     <- scale_sd(penguins_na.rm[, 1:4])
clas    <- penguins_na.rm$species
bas     <- basis_pca(dat)
mv      <- manip_var_of(bas)
mt_path <- manual_tour(bas, manip_var = mv)

## d = 2 case
message("facet_wrap_tour wants be called early, so that other proto's adopt the facet_var.")
ggt <- ggtour(mt_path, dat, angle = .3) +
  facet_wrap_tour(facet_var = clas, ncol = 2, nrow = 2) +
  proto_default(aes_args = list(color = clas, shape = clas),
                identity_args = list(size = 1.5))
## Not run: 
animate_gganimate(ggt) ## May not always play well with plotly

## End(Not run)

Description

Appends facet_wrap(vars(frame_number)) & minor themes to the ggtour. If the number of frames is more than desired, try increasing the angle argument on the tour. Is very demanding on the plots pane, works better with ggsave().

Usage

filmstrip(ggtour, ...)

Arguments

See Also

Other ggtour animator: animate_gganimate(), animate_plotly()

Examples

library(spinifex)
dat  <- scale_sd(penguins_na.rm[, 1:4])
clas <- penguins_na.rm$species
bas  <- basis_pca(dat)
mv   <- manip_var_of(bas)

## d = 2 case
mt_path <- manual_tour(bas, manip_var = mv)
ggt <- ggtour(mt_path, dat, angle = .3) +
  proto_point(list(color = clas, shape = clas),
              list(size = 1.5)) +
  proto_basis()
filmstrip(ggt)

## d = 1 case & specify facet dim
bas1d     <- basis_pca(dat, d = 1)
mt_path1d <- manual_tour(basis = bas1d, manip_var = mv)
ggt1d <- ggtour(mt_path1d, dat, angle = 99) +
  proto_default1d(aes_args = list(fill = clas, color = clas))
filmstrip(ggt1d, nrow = 12, ncol = 3)

Description

ggtour() initializes a ggplot object for a tour. ⁠proto_*⁠ functions are added to the tour, analogous to ⁠ggplot() + geom_*⁠. The final tour object is then animated with animate_plotly() or animate_ggtour(), or passed to filmstrip() for static plot faceting on frames.

Usage

ggtour(
  basis_array,
  data = NULL,
  angle = 0.05,
  basis_label = NULL,
  data_label = NULL,
  do_center_frame = TRUE
)

Arguments

See Also

Other ggtour proto functions: append_fixed_y(), facet_wrap_tour(), proto_basis(), proto_default(), proto_density(), proto_density2d(), proto_hex(), proto_highlight(), proto_hline0(), proto_origin(), proto_point(), proto_text()

Examples

library(spinifex)
dat     <- scale_sd(penguins_na.rm[, 1:4])
clas    <- penguins_na.rm$species
bas     <- basis_pca(dat)
mv      <- manip_var_of(bas)
mt_path <- manual_tour(bas, manip_var = mv)

## d = 2 case
ggt <- ggtour(basis_array = mt_path, data = dat, angle = .3) +
  proto_default(aes_args = list(color = clas, shape = clas),
                identity_args = list(size = 1.5, alpha = .8))

animate_plotly(ggt)


## Finer control calling individual proto_* functions
ggt <- ggtour(basis_array = mt_path, data = dat, angle = .3) +
  proto_point(aes_args = list(color = clas, shape = clas),
              identity_args = list(size = 1.5, alpha = .8),
              row_index = which(clas == levels(clas)[1])) +
  proto_basis(position = "right",
              manip_col = "red",
              text_size = 7L) +
  proto_origin()

animate_plotly(ggt)


## d = 1 case
bas1d <- basis_pca(dat, d = 1)
mt_path1d <- manual_tour(basis = bas1d, manip_var = mv)

ggt1d <- ggtour(basis_array = mt_path1d, data = dat, angle = .3) +
  proto_default1d(aes_args = list(fill= clas, color = clas))

animate_plotly(ggt1d)


## Single basis
ggt <- ggtour(basis_array = bas, data = dat) +
  proto_default(aes_args = list(fill= clas, color = clas))
## ggtour() returns a static ggplot2 plot

ggt
### or as html widget with tooltips
animate_plotly(ggt)

Description

Internal function. Interpolates a manual tour over the stored theta, and phi specifications. Returns an interpolated basis_array to be consumed by array2df.

Usage

interpolate_manual_tour(basis_array, angle = 0.05)

Arguments

See Also

Other manual tour adjacent functions: create_manip_space(), manip_var_of(), manual_tour(), rotate_manip_space()

Examples

## This function is not meant for external use
dat  <- scale_sd(wine[, 2:6])
clas <- wine$Type
bas  <- basis_pca(dat)
mv   <- manip_var_of(bas)
mt   <- manual_tour(bas, mv)

interp <- spinifex:::interpolate_manual_tour(basis_array = mt, angle = .1)
dim(interp)
str(interp)

Description

Checks if any of the layers of a ggplot contain a specific class.

Usage

is_any_layer_class(ggplot, class_nm = "GeomDensity")

Arguments

See Also

ggplot2::theme for all theme options.

Examples

library(ggplot2)
library(spinifex)

g <- ggplot(mtcars, aes(disp, color = factor(cyl))) + 
  geom_density() + geom_histogram()
is_any_layer_class(g, "GeomDensity")
is_any_layer_class(g, "GeomPoint")

Description

Test if a numeric matrix is orthonormal, that is, each column is orthogonal, at a right angle with the others, and each column has a norm length of 1. This must be true for a projection to be linear.

Usage

is_orthonormal(x, tol = 0.001)

Arguments

Value

Single logical, whether or not the matrix is orthonormal.

Examples

spinifex::is_orthonormal(tourr::basis_random(n = 6))
spinifex::is_orthonormal(matrix(1:12, ncol = 2), tol = 0.01)

Description

Find the column number of the variable with the rank-ith largest contribution of the basis. Useful for identifying a variable to change the contribution of in a manual tour, it's manip_var argument.

Usage

manip_var_of(basis, rank = 1)

Arguments

Value

Numeric scalar, the column number of a variable.

See Also

Other manual tour adjacent functions: create_manip_space(), interpolate_manual_tour(), manual_tour(), rotate_manip_space()

Examples

dat <- scale_sd(wine[, 2:6])
bas <- basis_pca(dat)

manip_var_of(basis = bas) ## Variable with the largest contribution
manip_var_of(basis = bas, rank = 5) ## Variable with 5th-largest contribution

Description

Typically called by array2af(). An array of projections, the radial tour of the manip_var, which is rotated from phi's starting position to phi_max, to phi_min, and back to the start position.

Usage

manual_tour(
  basis,
  manip_var,
  theta = NULL,
  phi_min = 0,
  phi_max = pi/2,
  data = NULL
)

Arguments

Value

A (p, d, 4) history_array of the radial tour. The bases set for phi_start, phi_min, phi_max, and back to phi_start.

See Also

Other manual tour adjacent functions: create_manip_space(), interpolate_manual_tour(), manip_var_of(), rotate_manip_space()

Examples

library(spinifex)
dat  <- scale_sd(wine[, 2:6])
clas <- wine$Type
bas  <- basis_pca(dat)
mv   <- manip_var_of(bas)
manual_tour(basis = bas, manip_var = mv)

## All arguments
manual_tour(basis = bas, manip_var = mv,
            theta = pi / 2, phi_min = pi / 16, phi_max = pi)

## Animating with ggtour() & proto_* (d = 2 case)
mt <- manual_tour(basis = bas, manip_var = mv)
ggt <- ggtour(mt, dat, angle = .2) +
  proto_origin() +
  proto_point(list(color = clas, shape = clas)) +
  proto_basis()

animate_plotly(ggt)


## d = 1 case
## basis could be 1- or 2D; protos_* only use 1st column
mv  <- manip_var_of(bas)
mt  <- manual_tour(basis = bas, manip_var = mv)
ggt <- ggtour(mt, dat, angle = .3) +
  proto_density(aes_args = list(color = clas, fill = clas)) +
  proto_basis1d() +
  proto_origin1d()

animate_plotly(ggt)


## Bring your own basis
bas <- matrix(rnorm(2 * ncol(dat)), ncol = 2)
bas <- orthonormalise(bas) ## manual_tour warns if basis isn't orthonormal
mt  <- manual_tour(basis = bas, manip_var = 1)
ggt <- ggtour(mt, dat, angle = .2) +
  proto_default(aes_args = list(color = clas, shape = clas))

animate_plotly(ggt)

Description

A manual variant of map_relative(). Can be used as the axes argument to manually set the size and locations of the axes.

Usage

map_absolute(x, offset = c(0, 0), scale = c(1, 1))

Arguments

Value

Scaled and offset x.

See Also

scale_axes for preset choices.

Other linear mapping functions: map_relative()

Examples

bas <- tourr::basis_random(4, 2)

map_absolute(bas, offset = c(-2, 0), scale = c(2/3, 2/3))

Description

Internal function. Typically called by other functions to scale the position of the axes data.frame or another data.frame to plot relative to the data.

Usage

map_relative(
  x,
  position = c("center", "left", "right", "bottomleft", "topright", "off", "top1d",
    "floor1d", "bottom1d", "full", "facetleft", "facetright", "facettop", "facetbottom"),
  to = NULL
)

Arguments

Value

Transformed values of x, dimension and class unchanged.

See Also

map_absolute for more manual control.

Other linear mapping functions: map_absolute()

Examples

## !!This function is not meant for external use!!
rb <- tourr::basis_random(4, 2)

map_relative(x = rb, position = "bottomleft")
map_relative(x = rb, position = "right", to = wine[, 2:3])

Description

Includes measurements for penguin species, island in Palmer Archipelago, size (flipper length, body mass, bill dimensions), and sex.

Usage

penguins_na.rm

Format

A data frame with 333 rows and 4 numeric variables and 3 factor variables

bill_length_mm

a number denoting bill length (millimeters)

bill_depth_mm

a number denoting bill depth (millimeters)

flipper_length_mm

an integer denoting flipper length (millimeters)

body_mass_g

an integer denoting body mass (grams)

species

a factor denoting penguin species (Adelie, Chinstrap and Gentoo)

sex

a factor denoting penguin sex (female, male)

island

a factor denoting island in Palmer Archipelago, Antarctica (Biscoe, Dream or Torgersen)

Details

This is a cleaned subset of palmerpenguins::penguins.

Replicating this dataset:

require(palmerpenguins)
d <- palmerpenguins::penguins
d <- d[complete.cases(d), ] ## Remove missing, 2 obs of numeric and several in sex
d <- d[, c(3:6, 1, 7, 2)]   ## Numeric to front, group factors, remove year
penguins_na.rm <- as.data.frame(d) ## Remove {tibble} dependency
## save(penguins_na.rm, file = "./data/penguins_na.rm.rda")

Source

palmerpenguins R package. A. Horst, 2020. Palmer Archipelago (Antarctica) Penguin Data. https://CRAN.R-project.org/package=palmerpenguins

Adelie penguins: Palmer Station Antarctica LTER and K. Gorman. 2020. Structural size measurements and isotopic signatures of foraging among adult male and female Adelie penguins (Pygoscelis adeliae) nesting along the Palmer Archipelago near Palmer Station, 2007-2009 ver 5. Environmental Data Initiative doi:10.6073/pasta/98b16d7d563f265cb52372c8ca99e60f

Gentoo penguins: Palmer Station Antarctica LTER and K. Gorman. 2020. Structural size measurements and isotopic signatures of foraging among adult male and female Gentoo penguin (Pygoscelis papua) nesting along the Palmer Archipelago near Palmer Station, 2007-2009 ver 5. Environmental Data Initiative doi:10.6073/pasta/7fca67fb28d56ee2ffa3d9370ebda689

Chinstrap penguins: Palmer Station Antarctica LTER and K. Gorman. 2020. Structural size measurements and isotopic signatures of foraging among adult male and female Chinstrap penguin (Pygoscelis antarcticus) nesting along the Palmer Archipelago near Palmer Station, 2007-2009 ver 6. Environmental Data Initiative doi:10.6073/pasta/c14dfcfada8ea13a17536e73eb6fbe9e

Originally published in: Gorman KB, Williams TD, Fraser WR (2014) Ecological Sexual Dimorphism and Environmental Variability within a Community of Antarctic Penguins (Genus Pygoscelis). PLoS ONE 9(3): e90081. doi:10.1371/journal.pone.0090081

Examples

library(spinifex)
str(penguins_na.rm)
dat   <- scale_sd(penguins_na.rm[, 1:4])
clas1 <- penguins_na.rm$species
clas2 <- penguins_na.rm$sex

bas <- basis_pca(dat)
mv  <- manip_var_of(bas)
mt  <- manual_tour(bas, mv)

ggt <- ggtour(mt, dat, angle = .2) +
  proto_default(aes_args = list(color = clas1, shape = clas2))

animate_plotly(ggt)

Description

The data set PimaIndiansDiabetes2 contains a corrected version of the original data set. While the UCI repository index claims that there are no missing values, closer inspection of the data shows several physical impossibilities, e.g., blood pressure or body mass index of 0. In PimaIndiansDiabetes2, all zero values of glucose, pressure, triceps, insulin and mass have been set to NA, see also Wahba et al (1995) and Ripley (1996).

Usage

PimaIndiansDiabetes_long

Format

A data frame with 724 observations of 6 numeric variables, and target factor diabetes.

• pregnant, Number of times pregnant

• glucose, Plasma glucose concentration (glucose tolerance test)

• pressure, Diastolic blood pressure (mm Hg)

• mass, Body mass index (weight in kg/(height in m, squared))

• pedigree, Diabetes pedigree function

• age, Age (years)

• diabetes, Class variable (test for diabetes), either "pos" or "neg"

Details

This is a cleaned subset of ⁠mlbench's⁠ PimaIndiansDiabetes2. See help(PimaIndiansDiabetes2, package = "mlbench").

Replicating this dataset:

require("mlbench")
data(PimaIndiansDiabetes2)

d <- PimaIndiansDiabetes2
d <- d[, c(1:3, 6:9)] ## Remove 2 colulmns with the most NAs
d <- d[complete.cases(d), ] ## Remove ~44 row-wise incomplete rows
PimaIndiansDiabetes_long <- d
## save(PimaIndiansDiabetes_long, file = "./data/PimaIndiansDiabetes_long.rda")

Source

J.W. Smith., el al. 1988. Using the ADAP learning algorithm to forecast the onset of diabetes mellitus. In Proceedings of the Symposium on Computer Applications and Medical Care (pp. 261–265). IEEE Computer Society Press.

mlbench, R package. F. Leisch & E. Dimitriadou, 2021. mlbench: Machine Learning Benchmark Problems https://CRAN.R-project.org/package=mlbench

Examples

library(spinifex)
str(PimaIndiansDiabetes_long)
dat  <- scale_sd(PimaIndiansDiabetes_long[, 1:6])
clas <- PimaIndiansDiabetes_long$diabetes

bas <- basis_pca(dat)
mv  <- manip_var_of(bas)
mt  <- manual_tour(bas, mv)

ggt <- ggtour(mt, dat, angle = .2) +
  proto_default(aes_args = list(color = clas, shape = clas))

animate_plotly(ggt)

Description

The data set PimaIndiansDiabetes2 contains a corrected version of the original data set. While the UCI repository index claims that there are no missing values, closer inspection of the data shows several physical impossibilities, e.g., blood pressure or body mass index of 0. In PimaIndiansDiabetes2, all zero values of glucose, pressure, triceps, insulin and mass have been set to NA, see also Wahba et al (1995) and Ripley (1996).

Usage

PimaIndiansDiabetes_wide

Format

A data frame with 392 observations of 8 numeric variables, and target factor diabetes.

• pregnant, Number of times pregnant

• glucose, Plasma glucose concentration (glucose tolerance test)

• pressure, Diastolic blood pressure (mm Hg)

• triceps, Triceps skin fold thickness (mm)

• insulin, 2-Hour serum insulin (mu U/ml)

• mass, Body mass index (weight in kg/(height in m, squared))

• pedigree, Diabetes pedigree function

• age, Age (years)

• diabetes, Class variable (test for diabetes), either "pos" or "neg"

Details

This is a cleaned subset of mlbench's PimaIndiansDiabetes2. See help(PimaIndiansDiabetes2, package = "mlbench").

Replicating this dataset:

require("mlbench")
data(PimaIndiansDiabetes2)

d <- PimaIndiansDiabetes2
d <- d[complete.cases(d), ] ## Remove ~350 row-wise incomplete rows
PimaIndiansDiabetes_wide <- d
## save(PimaIndiansDiabetes_wide, file = "./data/PimaIndiansDiabetes_wide.rda")

Examples

library(spinifex)
str(PimaIndiansDiabetes_wide)
dat  <- scale_sd(PimaIndiansDiabetes_wide[, 1:8])
clas <- PimaIndiansDiabetes_wide$diabetes

bas <- basis_pca(dat)
mv  <- manip_var_of(bas)
mt  <- manual_tour(bas, mv)

ggt <- ggtour(mt, dat, angle = .2) +
  proto_default(aes_args = list(color = clas, shape = clas))

animate_plotly(ggt)

Description

, see ggtour. Performs the a manual tour and returns an animation of render_type. For use with tourr::save_history() tour paths see play_tour_path().

Usage

play_manual_tour(
  basis = NULL,
  data,
  manip_var,
  theta = NULL,
  phi_min = 0,
  phi_max = 0.5 * pi,
  angle = 0.05,
  render_type = render_plotly,
  ...
)

Arguments

Value

An animation of a radial tour.

See Also

render_ For arguments to pass into ....

Examples

library(spinifex)
message("It's suggested to switch to the proto api, see `?ggtour` to get started.")

## Setup
dat_std <- scale_sd(wine[, 2:6])
clas    <- wine$Type
bas     <- basis_pca(dat_std)
mv      <- manip_var_of(bas)

## Not run: 
suppressWarnings(
  play_manual_tour(basis = bas, data = dat_std, manip_var = mv)
)

suppressWarnings(
  play_manual_tour(
    basis = bas, data = dat_std, manip_var = mv,
    theta = .5 * pi, axes = "right", fps = 5,
    angle = .08, phi_min = 0, phi_max = 2 * pi,
    aes_args = list(color = clas, shape = clas),
    identity_args = list(size = 1.5, alpha = .7),
    ggproto = list(ggplot2::theme_void(), ggplot2::ggtitle("My title")),
    render_type = render_gganimate)
)
## Saving output may require additional setup
if(FALSE){ ## Don't accidentally save file
  ## Export plotly .html widget
  play_manual_tour(basis = bas, data = dat_std, manip_var = 6,
                   render_type = render_plotly,
                   html_filename = "myRadialTour.html")
  
  ## Export gganimate .gif
  play_manual_tour(basis = bas, data = dat_std, manip_var = 1,
                   render_type = render_gganimate,
                   gif_filename = "myRadialTour.gif", gif_path = "./output")
}

## End(Not run)

Description

, see ggtour. Takes the result of tourr::save_history() or manual_tour(), interpolates over the path and renders into a specified render_type.

Usage

play_tour_path(
  tour_path,
  data = NULL,
  angle = 0.05,
  render_type = render_plotly,
  ...
)

Arguments

See Also

render_ For arguments to pass into ....

Examples

library(spinifex)
message("It's suggested to switch to the proto api, see `?ggtour` to get started.")

dat_std <- scale_sd(wine[, 2:6])
clas    <- wine$Type
bas     <- basis_pca(dat_std)
gt_path <- save_history(dat_std, tour_path = tourr::grand_tour(), max = 5)

## Not run: 
suppressWarnings(
  play_tour_path(tour_path = gt_path, data = dat_std)
)

suppressWarnings(
  play_tour_path(tour_path = gt_path, data = dat_std,
                 axes = "bottomleft", angle = .08, fps = 8,
                 aes_args = list(color = clas, shape = clas),
                 identity_args = list(size = 1.5, alpha = .7),
                 ggproto = 
                   list(ggplot2::theme_void(), ggplot2::ggtitle("My title")),
                 render_type = render_gganimate)
)

## Saving a .gif(may require additional setup)
if(FALSE){ ## Don't accidentally save file
  ## Export plotly .html widget
  play_tour_path(tour_path = gt_path, data = dat_std,
                 render_type = render_plotly,
                 html_filename = "myRadialTour.html")
               
  ## Export gganimate .gif
  play_tour_path(tour_path = gt_path, data = dat_std,
                 render_type = render_gganimate,
                 gif_path = "myOutput", gif_filename = "myRadialTour.gif")
}

## End(Not run)

Description

Adds basis axes to the animation, the direction and magnitude of contributions of the variables to the projection space inscribed in a unit circle for 2D or rectangle of unit width for 1D.

Usage

proto_basis(
  position = c("left", "center", "right", "bottomleft", "topright", "full", "off"),
  manip_col = "blue",
  line_size = 0.6,
  text_size = 4
)

proto_basis1d(
  position = c("bottom1d", "floor1d", "top1d", "full", "off"),
  manip_col = "blue",
  segment_size = 2,
  text_size = 4,
  text_offset = -1.15
)

Arguments

See Also

Other ggtour proto functions: append_fixed_y(), facet_wrap_tour(), ggtour(), proto_default(), proto_density(), proto_density2d(), proto_hex(), proto_highlight(), proto_hline0(), proto_origin(), proto_point(), proto_text()

Examples

library(spinifex)
dat  <- scale_sd(penguins_na.rm[, 1:4])
clas <- penguins_na.rm$species
bas  <- basis_pca(dat)
mv   <- manip_var_of(bas)

## 2D case:
mt_path <- manual_tour(bas, manip_var = mv)
ggt <- ggtour(mt_path, dat, angle = .3) +
  proto_point() +
  proto_basis()

animate_plotly(ggt)


## Customize basis
ggt2 <- ggtour(mt_path, dat) +
  proto_basis(position = "right", manip_col = "green",
              line_size = .8, text_size = 8)

animate_plotly(ggt2)


## 1D case:
bas1d     <- basis_pca(dat, d = 1)
mv        <- manip_var_of(bas1d, 3)
mt_path1d <- manual_tour(bas1d, manip_var = mv)

ggt1d <- ggtour(mt_path1d, dat, angle = .3) +
  proto_density() +
  proto_basis1d()

animate_plotly(ggt1d)


## Customized basis1d
ggt1d <- ggtour(mt_path1d, dat, angle = .3) +
  proto_density() +
  proto_basis1d(position     = "bottom",
                manip_col    = "pink",
                segment_size = 3,
                text_size    = 6,
                text_offset  = 1.2)

animate_plotly(ggt1d)

Description

An easier way to get to default 2D tour settings. Returns a list of proto_origin(), proto_point(...), proto_basis() for 2D. Returns a list of proto_origin1d(), proto_density(...), proto_basis1d() for 1D.

Usage

proto_default(
  position = c("left", "center", "right", "bottomleft", "topright", "off"),
  ...
)

proto_default1d(position = c("bottom1d", "floor1d", "top1d", "off"), ...)

Arguments

See Also

Other ggtour proto functions: append_fixed_y(), facet_wrap_tour(), ggtour(), proto_basis(), proto_density(), proto_density2d(), proto_hex(), proto_highlight(), proto_hline0(), proto_origin(), proto_point(), proto_text()

Examples

library(spinifex)
dat  <- scale_sd(penguins_na.rm[, 1:4])
clas <- penguins_na.rm$species

## 2D case:
bas     <- basis_pca(dat)
mv      <- manip_var_of(bas)
mt_path <- manual_tour(bas, mv)

ggt <- ggtour(mt_path, dat) +
  proto_default(aes_args = list(color = clas, shape = clas))

animate_plotly(ggt)

library(spinifex)

## 1D case:
gt_path <- save_history(dat, grand_tour(d = 1), max_bases = 3)

ggt <- ggtour(gt_path, dat) +
  proto_default1d(aes_args = list(fill = clas, color = clas))

animate_plotly(ggt)

Description

Adds geom_density() and geom_rug() of the projected data. Density postion = "stack" does not work with animate_plotly(), GH issue is open.

Usage

proto_density(
  aes_args = list(),
  identity_args = list(alpha = 0.7),
  row_index = NULL,
  density_position = c("identity", "stack", "fill"),
  rug_shape = c(3, 142, 124, NULL)
)

Arguments

See Also

Other ggtour proto functions: append_fixed_y(), facet_wrap_tour(), ggtour(), proto_basis(), proto_default(), proto_density2d(), proto_hex(), proto_highlight(), proto_hline0(), proto_origin(), proto_point(), proto_text()

Examples

library(spinifex)
dat  <- scale_sd(penguins_na.rm[, 1:4])
clas <- penguins_na.rm$species

## Manual tour
bas <- basis_olda(dat, clas)
mt  <- manual_tour(bas, manip_var = 2)
ggt <- ggtour(mt, dat, angle = .3) +
  proto_density(aes_args = list(color = clas, fill = clas)) +
  proto_basis1d() +
  proto_origin1d()

animate_plotly(ggt)


## Grand tour
gt_path <- save_history(dat, grand_tour(), max = 3)
ggt <- ggtour(gt_path, dat, angle = .3) +
  proto_density(aes_args = list(color = clas, fill = clas)) +
  proto_basis1d() +
  proto_origin1d()

animate_plotly(ggt)

Description

Adds geom_density_2d() of the projected data.

Usage

proto_density2d(
  aes_args = list(),
  identity_args = list(bins = 4),
  row_index = NULL
)

Arguments

See Also

Other ggtour proto functions: append_fixed_y(), facet_wrap_tour(), ggtour(), proto_basis(), proto_default(), proto_density(), proto_hex(), proto_highlight(), proto_hline0(), proto_origin(), proto_point(), proto_text()

Examples

library(spinifex)
dat     <- scale_sd(penguins_na.rm[, 1:4])
clas    <- penguins_na.rm$species
gt_path <- save_history(dat, grand_tour(), max = 3)

## geom_density_2d args can be passed in identity_args (bins, binwidth, breaks) 
ggt <- ggtour(gt_path, dat, angle = .3) +
  proto_density2d(aes_args = list(color = clas, fill = clas),
                  identity_args = list(binwidth = .3)) +
  proto_point(aes_args = list(color = clas, shape = clas),
              identity_args = list(alpha = .2)) +
  proto_basis()

animate_plotly(ggt)

Description

Adds text to the animation, the frame and its specified correlation.

Usage

proto_frame_cor2(
  text_size = 4,
  row_index = TRUE,
  xy_position = c(0.7, -0.1),
  ...
)

Arguments

See Also

stats::cor

Examples

library(spinifex)
dat     <- scale_sd(penguins_na.rm[, 1:4])
clas    <- penguins_na.rm$species
gt_path <- save_history(dat, grand_tour(), max_bases = 5)

ggt <- ggtour(gt_path, dat, angle = .3) +
  proto_default(aes_args = list(color = clas, shape = clas)) +
  proto_frame_cor2(xy_position = c(.5, 1.1))

animate_plotly(ggt)

Description

Adds geom_hex() of the projected data. Does not display hexagons in plotly animations; will not work with animate_plotly().

Usage

proto_hex(
  aes_args = list(),
  identity_args = list(),
  row_index = NULL,
  bins = 30
)

Arguments

See Also

Other ggtour proto functions: append_fixed_y(), facet_wrap_tour(), ggtour(), proto_basis(), proto_default(), proto_density(), proto_density2d(), proto_highlight(), proto_hline0(), proto_origin(), proto_point(), proto_text()

Examples

library(spinifex)
raw     <- ggplot2::diamonds
dat     <- scale_sd(raw[1:10000, c(1, 5:6, 8:10)])
gt_path <- save_history(dat, grand_tour(), max = 3)

## 10000 rows is quite heavy to animate.
## Increase performance by aggregating many points into few hexagons
ggp <- ggtour(gt_path, dat) +
  proto_basis() +
  proto_hex(bins = 20)

## Hexagons don't show up in plotly animation.
## Not run: 
animate_gganimate(ggp)

## End(Not run)

Description

A geom_point or geom_segment(1d case) call to draw attention to a subset of points. This is mostly redundant proto_point with the implementation of the row_index argument on data protos, still helpful in the 1d case and for mark_initial, does not use bkg_row_color

Usage

proto_highlight(
  aes_args = list(),
  identity_args = list(color = "red", size = 5, shape = 8),
  row_index = 1,
  mark_initial = FALSE
)

proto_highlight1d(
  aes_args = list(),
  identity_args = list(color = "red", linetype = 2, alpha = 0.9),
  row_index = 1,
  mark_initial = FALSE
)

Arguments

See Also

Other ggtour proto functions: append_fixed_y(), facet_wrap_tour(), ggtour(), proto_basis(), proto_default(), proto_density(), proto_density2d(), proto_hex(), proto_hline0(), proto_origin(), proto_point(), proto_text()

Examples

library(spinifex)
dat     <- scale_sd(penguins_na.rm[, 1:4])
clas    <- penguins_na.rm$species
gt_path <- save_history(dat, grand_tour(), max_bases = 5)

## d = 2 case
ggt <- ggtour(gt_path, dat, angle = .3) +
  proto_default(aes_args = list(color = clas, shape = clas)) +
  proto_highlight(row_index = 5)

animate_plotly(ggt)


## Highlight multiple observations
ggt2 <- ggtour(gt_path, dat, angle = .3) +
  proto_default(aes_args = list(color = clas, shape = clas)) +
  proto_highlight(row_index = c( 2, 6, 19),
                  identity_args = list(color = "blue", size = 4, shape = 4))

animate_plotly(ggt2)

## 1D case:
gt_path1d <- save_history(dat, grand_tour(d = 1), max_bases = 3)

ggt <- ggtour(gt_path1d, dat, angle = .3) +
  proto_default1d(aes_args = list(fill = clas, color = clas)) +
  proto_highlight1d(row_index = 7)

animate_plotly(ggt)


## Highlight multiple observations, mark_initial defaults to off
ggt2 <- ggtour(gt_path1d, dat, angle = .3) +
  proto_default1d(aes_args = list(fill = clas, color = clas)) +
  proto_highlight1d(row_index = c(2, 6, 7),
                    identity_args = list(color = "green", linetype = 1))

animate_plotly(ggt2)

Description

Adds a vertical/horizontal line with an intercept of 0, scaled to the data frame.

Usage

proto_hline0(identity_args = list(color = "grey80", size = 0.5, alpha = 0.9))

proto_vline0(identity_args = list(color = "grey80", size = 0.5, alpha = 0.9))

Arguments

See Also

Other ggtour proto functions: append_fixed_y(), facet_wrap_tour(), ggtour(), proto_basis(), proto_default(), proto_density(), proto_density2d(), proto_hex(), proto_highlight(), proto_origin(), proto_point(), proto_text()

Examples

library(spinifex)
dat  <- scale_sd(penguins_na.rm[, 1:4])
clas <- penguins_na.rm$species

## 2D case:
gt_path <- save_history(dat, grand_tour(), max_bases = 5)
ggt <- ggtour(gt_path, dat, angle = .1) +
  proto_point(list(color = clas, shape = clas)) +
  proto_hline0() + ## horizonatal line at 0
  proto_vline0()   ## vertical line at 0

animate_plotly(ggt)

Description

Adds a zero mark showing the location of the origin for the central data area.

Usage

proto_origin(
  identity_args = list(color = "grey60", size = 0.5, alpha = 0.9),
  tail_size = 0.05
)

proto_origin1d(identity_args = list(color = "grey60", size = 0.5, alpha = 0.9))

Arguments

See Also

Other ggtour proto functions: append_fixed_y(), facet_wrap_tour(), ggtour(), proto_basis(), proto_default(), proto_density(), proto_density2d(), proto_hex(), proto_highlight(), proto_hline0(), proto_point(), proto_text()

Examples

library(spinifex)
dat  <- scale_sd(penguins_na.rm[, 1:4])
clas <- penguins_na.rm$species

## 2D case:
gt_path <- save_history(dat, grand_tour(), max_bases = 5)
ggt <- ggtour(gt_path, dat, angle = .1) +
  proto_point(list(color = clas, shape = clas)) +
  proto_origin() ## `+` in center

animate_plotly(ggt)


## 1D case:
gt_path1d <- save_history(dat, grand_tour(d = 1), max_bases = 5)

ggt <- ggtour(gt_path1d, dat) +
  proto_density(list(fill = clas, color = clas)) +
  proto_origin1d() ## Adds line at 0.

animate_plotly(ggt)

Description

Adds geom_point() of the projected data.

Usage

proto_point(
  aes_args = list(),
  identity_args = list(alpha = 0.9),
  row_index = NULL,
  bkg_color = "grey80"
)

Arguments

See Also

Other ggtour proto functions: append_fixed_y(), facet_wrap_tour(), ggtour(), proto_basis(), proto_default(), proto_density(), proto_density2d(), proto_hex(), proto_highlight(), proto_hline0(), proto_origin(), proto_text()

Examples

library(spinifex)
dat     <- scale_sd(penguins_na.rm[, 1:4])
clas    <- penguins_na.rm$species
gt_path <- save_history(dat, grand_tour(), max_bases = 5)

ggt <- ggtour(gt_path, dat, angle = .3) +
  proto_point(aes_args = list(color = clas, shape = clas),
              identity_args = list(size = 2, alpha = .7))

animate_plotly(ggt)


## Select/highlight observations with `row_index`
ggt <- ggtour(gt_path, dat, angle = .3) +
  proto_point(aes_args = list(color = clas, shape = clas),
              identity_args = list(size = 2, alpha = .7),
              row_index = which(clas == levels(clas)[1]),
              bkg_color = "grey80") ## FALSE or NULL to skip plotting background

animate_plotly(ggt)

Description

Adds geom_text() of the projected data.

Usage

proto_text(
  aes_args = list(vjust = "outward", hjust = "outward"),
  identity_args = list(nudge_x = 0.05),
  row_index = TRUE
)

Arguments

See Also

Other ggtour proto functions: append_fixed_y(), facet_wrap_tour(), ggtour(), proto_basis(), proto_default(), proto_density(), proto_density2d(), proto_hex(), proto_highlight(), proto_hline0(), proto_origin(), proto_point()

Examples

library(spinifex)
dat     <- scale_sd(penguins_na.rm[, 1:4])
clas    <- penguins_na.rm$species
bas     <- basis_pca(dat)
mv      <- manip_var_of(bas)
gt_path <- save_history(dat, grand_tour(), max_bases = 5)

ggt <- ggtour(gt_path, dat, angle = .2) +
  proto_text(list(color = clas))

animate_plotly(ggt)


## Custom labels, subset of points
ggt2 <- ggtour(gt_path, dat) +
  proto_text(list(color = clas, size = as.integer(clas)),
             list(alpha = .7),
             row_index = 1:15)

animate_plotly(ggt2)

Description

, see ggtour. Typically called by render_plotly() or render_gganimate(). Takes the result of array2df(), and renders them into a ggplot2 object.

Usage

render_(
  frames,
  axes = "center",
  manip_col = "blue",
  line_size = 0.6,
  text_size = 4,
  aes_args = list(),
  identity_args = list(),
  ggproto = list(theme_spinifex())
)

Arguments

Examples

library(spinifex)
message("It's suggested to switch to the proto api, see `?ggtour` to get started.")

## Setup
dat_std <- scale_sd(wine[, 2:6])
clas    <- wine$Type
bas     <- basis_pca(dat_std)
mv      <- manip_var_of(bas)

mt_array <- manual_tour(basis = bas, manip_var = mv)
mt_df_ls <- array2df(basis_array = mt_array, data = dat_std)

## Required arguments
render_(frames = mt_df_ls)

## Full arguments
require("ggplot2")
render_(frames = mt_df_ls, axes = "left", manip_col = "purple",
        aes_args = list(color = clas, shape = clas),
        identity_args = list(size = 1.5, alpha = .7),
        ggproto = list(theme_minimal(),
                       ggtitle("My title"),
                       scale_color_brewer(palette = "Set2")))

Description

, see ggtour. Takes the result of array2df() and renders them into a gganimate animation.

Usage

render_gganimate(
  fps = 8,
  rewind = FALSE,
  start_pause = 0.5,
  end_pause = 1,
  gif_filename = NULL,
  gif_path = NULL,
  gganimate_args = list(),
  ...
)

Arguments

See Also

render_ for ... arguments.

gganimate::anim_save for more control of .gif output.

Examples

library(spinifex)
message("It's suggested to switch to the proto api, see `?ggtour` to get started.")

## Setup
dat_std  <- scale_sd(wine[, 2:6])
clas     <- wine$Type
bas      <- basis_pca(dat_std)
mv       <- manip_var_of(bas)
mt       <- manual_tour(basis = bas, manip_var = mv)
mt_df_ls <- array2df(basis_array = mt, data = dat_std)

## Not run: 
render_gganimate(frames = mt_df_ls)

require("ggplot2")
render_gganimate(
  frames = mt_df_ls, axes = "bottomleft",
  fps = 10, rewind = TRUE, start_pause = 1, end_pause = 1.5,
  aes_args = list(color = clas, shape = clas),
  identity_args = list(size = 2, alpha = .7),
  ggproto = list(theme_void(),
                 ggtitle("My title"),
                 scale_color_brewer(palette = "Set2")))

## Saving a .gif(may require additional setup)
if(FALSE) ## Don't run by mistake
  render_gganimate(frames = mt_df_ls, axes = "bottomleft",
                   gif_filename = "myRadialTour.gif", gif_path = "./output")
## End(Not run)

Description

, see ggtour. Takes the result of array2df() and animations them via {plotly} into a self-contained HTML widget.

Usage

render_plotly(fps = 8, html_filename = NULL, save_widget_args = list(), ...)

Arguments

See Also

render_ for ... arguments.

ggplotly for source documentation of tooltip.

saveWidget for more control of .html output.

Examples

library(spinifex)
message("It's suggested to switch to the proto api, see `?ggtour` to get started.")

## Setup
dat_std  <- scale_sd(wine[, 2:6])
clas     <- wine$Type
bas      <- basis_pca(dat_std)
mv       <- manip_var_of(bas)
mt_array <- manual_tour(basis = bas, manip_var = mv)
mt_df_ls <- array2df(basis_array = mt_array, data = dat_std)


render_plotly(frames = mt_df_ls)

require("ggplot2")
render_plotly(
  frames = mt_df_ls, axes = "bottomleft", fps = 10,
  aes_args = list(color = clas, shape = clas),
  identity_args = list(size = 1.5, alpha = .7),
  ggproto = list(theme_bw(), scale_color_brewer(palette = "Set2")))

## Saving a .gif, may require additional setup
if(FALSE) ## Don't accidentally save file
  render_plotly(frames = mt_df_ls, axes = "bottomleft", fps = 10,
                html_filename = "myRadialTour.html")

Description

A specific R3 rotation of the manipulation space for a 2D tour. Typically called by manual_tour(). The first 2 columns are x and y in the projection plane. The 3rd column extends "in the z-direction" orthogonal to the projection plane.

Usage

rotate_manip_space(manip_space, theta, phi)

Arguments

Value

A (p, d+1) orthonormal matrix of the rotated (manipulation) space. The first 2 columns are x and y in the projection plane. The 3rd column extends "in the z-direction" orthogonal to the projection plane.

See Also

Other manual tour adjacent functions: create_manip_space(), interpolate_manual_tour(), manip_var_of(), manual_tour()

Examples

library(spinifex)
dat <- scale_sd(wine[, 2:6])
bas <- basis_pca(dat)
mv  <- manip_var_of(bas)
msp <- create_manip_space(basis = bas, manip_var = mv)
rotate_manip_space(msp, theta = runif(1, max = 2 * pi),
                   phi = runif(1, max = 2 * pi))

## d = 1 case
bas1d <- basis_pca(dat, d = 1)
mv    <- manip_var_of(bas1d)
msp   <- create_manip_space(bas1d, mv)
rotate_manip_space(msp, theta = 0, phi = runif(1, max = 2 * pi))

Description

Runs a local shiny app that demonstrates manual tour and comparable traditional techniques for static projections of multivariate data sets.

Usage

run_app(app_nm = "radial_tour", ...)

Arguments

Value

Runs a locally hosted shiny app.

Examples

## Not run: 
run_app("radial_tour")
run_app(app_nm = "radial_tour", display.mode = "showcase")
## End(Not run)

Description

Save a tour path so it can later be displayed in many different ways. A wrapper function can mute the noisy text side effects of tourr::save_history. Changes a few argument defaults differ: doesn't scale data columns to (0, 1), max_bases = 10, appends the start basis if tour_path is grand, it isn't already there, and has correct dim.

Usage

save_history(
  data,
  tour_path = tourr::grand_tour(),
  max_bases = 10,
  start = NULL,
  rescale = FALSE,
  sphere = FALSE,
  step_size = Inf,
  verbose = getOption("verbose"),
  ...
)

Arguments

See Also

tourr::save_history tourr::new_tour tourr::grand_tour

Examples

library(spinifex)
dat <- scale_sd(penguins_na.rm[, 1:4])
## A grand tour path
gt_path <- save_history(data = dat, tour_path = grand_tour(), max_bases = 10)
dim(gt_path)

## A 1d grand tour path
gt1d_path <- save_history(dat, grand_tour(d = 1), 10)
dim(gt1d_path)

## A holes guided tour path
holes_path <- save_history(dat, guided_tour(holes(), max.tries = 10))
dim(holes_path)

## These are basis_arrays to be used in ?spinifex::ggtour()

Description

Masks ggplot2's default color/fill color palette for discrete variables.

Usage

scale_colour_discrete(...)

scale_fill_discrete(...)

Arguments

Description

Centers and scales each column by standard deviation (sd) or to the interval (0, 1).

Usage

scale_sd(data)

scale_01(data)

Arguments

Examples

scale_sd(data = wine[, 2:6])
scale_01(data = wine[, 2:6])

Description

spinifex is a package that extends the package tourr. It builds the functionality for manual tours and allows other tours to be rendered by plotly or gganimate. Tours are a class of dynamic linear (orthogonal) projections of numeric multivariate data from p down to d dimensions that are viewed as an animation as p-space is rotated. Manual tours manipulate a selected variable, exploring how they contribute to the sensitivity of the structure in the projection. This is particularly useful after finding an interesting basis, perhaps via a guided tour optimizing the projection for some objective function.

Details

GitHub: https://github.com/nspyrison/spinifex

Author(s)

Maintainer: Nicholas Spyrison [email protected] (ORCID)

Authors:

• Dianne Cook (ORCID) [thesis advisor]

See Also

manual_tour() ggtour() proto_default()

Description

A ggplot2 theme suggested for linear projections with spinifex. The default theme in spinifex functions.

Usage

theme_spinifex(...)

Arguments

See Also

ggplot2::theme for all theme options.

Examples

theme_spinifex()

require("ggplot2")
ggplot(mtcars, aes(wt, mpg, color = as.factor(cyl))) +
  geom_point() + theme_spinifex()

Description

, see ggtour. Projects the specified rotation as a 2D ggplot object. One static frame of manual tour. Useful for providing user-guided interaction.

Usage

view_frame(
  basis = NULL,
  data = NULL,
  manip_var = NULL,
  theta = 0,
  phi = 0,
  basis_label = abbreviate(row.names(basis), 3),
  rescale_data = FALSE,
  ...
)

Arguments

Value

A ggplot object of the rotated projection.

See Also

proto_default For arguments to pass into ....

Examples

library(spinifex)
message("It's suggested to switch to the proto api, see `?ggtour` to get started.")

## Setup
dat_std <- scale_sd(wine[, 2:6])
clas    <- wine$Type
bas     <- basis_pca(dat_std)
mv      <- manip_var_of(bas)

## Minimal example

suppressWarnings(
  view_frame(basis = bas)
)

## Typical example
suppressWarnings(
  view_frame(basis = bas, data = dat_std, manip_var = mv, axes = "left")
)

## Full example
rtheta <- runif(1, 0, 2 * pi)
rphi   <- runif(1, 0, 2 * pi)
suppressWarnings(
  view_frame(
    basis = bas, data = dat_std, manip_var = mv,
    theta = rtheta, phi = rphi, basis_label = paste0("MyNm", 1:ncol(dat_std)), 
    aes_args = list(color = clas, shape = clas),
    identity_args = list(size = 1.5, alpha = .7))
)

Description

Uses base graphics to plot the circle with axes representing the projection frame. Returns the corresponding table. Only works for 2d manual tours.

Usage

view_manip_space(
  basis,
  manip_var,
  tilt = 0.1 * pi,
  basis_label = abbreviate(row.names(basis), 3),
  manip_col = "blue",
  manip_sp_col = "red",
  line_size = 0.6,
  text_size = 4
)

Arguments

Value

ggplot object of the basis.

Examples

library(spinifex)

dat_std <- scale_sd(wine[, 2:6])
bas     <- basis_pca(dat_std)
mv      <- manip_var_of(bas)

view_manip_space(basis = bas, manip_var = mv)

view_manip_space(basis = bas, manip_var = mv,
                 tilt = 2/12 * pi, basis_label = paste0("MyNm", 1:ncol(dat_std)),
                 manip_col = "purple", manip_sp_col = "orange")

Description

One year of daily weather observations collected from the Canberra airport in Australia was obtained from the Australian Commonwealth Bureau of Meteorology and processed to create this sample dataset for illustrating data mining using R and Rattle.

Usage

weather_na.rm

Format

A data frame of 354 observations of 20 variables. One year of daily observations of weather variables at Canberra airport in Australia between November 1, 2007 and October 31, 2008.

• Date, The date of observation (Date class).

• MinTemp, The minimum temperature in degrees Celsius.

• MaxTemp, The maximum temperature in degrees Celsius.

• Rainfall, The amount of rainfall recorded for the day in mm.

• Evaporation, The "Class A pan evaporation" (mm) in the 24 hours to 9am.

• WindSpeed3pm, Wind speed (km/hr) averaged over 10 minutes prior to 3pm.

• Humid9am, Relative humidity (percent) at 9am.

• Humid3pm, Relative humidity (percent) at 3pm.

• Pressure9am, Atmospheric pressure (hpa) reduced to mean sea level at 9am.

• Pressure3pm, Atmospheric pressure (hpa) reduced to mean sea level at 3pm.

• Cloud9am, Fraction of sky obscured by cloud at 9am. This is measured in "oktas", which are a unit of eighths. It records how many eighths of the sky are obscured by cloud. A 0 measure indicates completely clear sky whilst an 8 indicates that it is completely overcast.

• Cloud3pm, Fraction of sky obscured by cloud (in "oktas": eighths) at 3pm. See Cloud9am for a description of the values.

• Temp9am, Temperature (degrees C) at 9am.

• Temp3pm, Temperature (degrees C) at 3pm.

• RISK_MM, The amount of rain. A kind of measure of the "risk".

• RainToday, Factor: "yes" if precipitation (mm) in the 24 hours to 9am exceeds 1mm, otherwise 0.

• RainTomorrow, Factor: "yes" if it rained the following day, the target variable.

Copyright Commonwealth of Australia 2010, Bureau of Meteorology. Definitions adapted from http://www.bom.gov.au/climate/dwo/IDCJDW0000.shtml

Details

The data has been processed to provide a target variable RainTomorrow (whether there is rain on the following day - No/Yes) and a risk variable RISK_MM (how much rain recorded in millimeters). Various transformations were performed on the source data. The dataset is quite small and is useful only for repeatable demonstration of various data science operations.

This is a cleaned subset of rattle::weather.

Replicating this dataset:

require("rattle")
d <- rattle::weather[, c(1, 3:7, 9, 12:21, 23, 22, 24)]
d <- d[complete.cases(d), ] ## Remove ~12 row-wise incomplete rows
d <- as.data.frame(d)  ## Remove tibble dependency
weather_na.rm <- d
## save(weather_na.rm, file = "./data/weather_na.rm.rda")

Source

Bureau of Meteorology, Commonwealth of Australia http://www.bom.gov.au/climate/data/

rattle, R package. G. Williams, 2020. rattle: Graphical User Interface for Data Science in R https://CRAN.R-project.org/package=rattle

Examples

library(spinifex)
str(weather_na.rm)
dat  <- scale_sd(weather_na.rm[, 2:18])
clas <- weather_na.rm$RainTomorrow

bas <- basis_pca(dat)
mv  <- manip_var_of(bas)
mt  <- manual_tour(bas, mv)

ggt <- ggtour(mt, dat, angle = .2) +
  proto_default(aes_args = list(color = clas, shape = clas))

animate_plotly(ggt)

Description

The wine dataset contains the results of a chemical analysis of wines grown in a specific area of Italy. Three types of wine are represented in the 178 samples, with the results of 13 chemical analyses recorded for each sample. The Type variable has been transformed into a categorical variable.

Usage

wine

Format

A data frame of 178 observations of target class Type and 12 numeric variables:

• Type, The type of wine, the target factor, 1 (59 obs), 2(71 obs), and 3 (48 obs).

• Alcohol, Alcohol

• Malic, Malic acid

• Ash, Ash

• Alcalinity, Alcalinity of ash

• Magnesium, Magnesium

• Phenols, Total phenols

• Flavanoids, Flavanoids

• Nonflavanoids, Nonflavanoid phenols

• Proanthocyanins, Proanthocyanins

• Color, Color intensity

• Hue, Hue

• Dilution, D280/OD315 of diluted wines

• Proline, Proline

Details

The data contains no missing values and consist of only numeric data, with a three class target variable (Type) for classification.

Replicating this dataset:

require("rattle")
str(rattle::wine)
## save(wine, file = "./data/wine.rda")

Source

rattle, R package. G. Williams, 2020. rattle: Graphical User Interface for Data Science in R https://CRAN.R-project.org/package=rattle

PARVUS. M. Forina. et al. 1988. Elsevier, Amsterdam, PARVUS: An extendable package of programs for data exploration, classification and correlation. ISBN 0-44-430121z

Examples

library(spinifex)
str(wine)
dat  <- scale_sd(wine[, 2:6])
clas <- wine$Type

bas <- basis_pca(dat)
mv  <- manip_var_of(bas)
mt  <- manual_tour(bas, mv)

ggt <- ggtour(mt, dat, angle = .2) +
  proto_default(aes_args = list(color = clas, shape = clas))

animate_plotly(ggt)