animate_paths {anipaths}

animate paths

Description

Animates telemetry data for the purposed of EDA using smoothing splines to interpolate the observed locations. The animations are particularly useful when examining multiple simultaneous trajectories. The output of the call to animate_paths() should bring up a browser window that shows the animation. Additionally, the images generated in images/ (or else the value set for imgdir) may be used with ffmpeg, latex, or other presentation software that can build animations directly from a sequence of images.

Usage

animate_paths(
  paths,
  coord = c("x", "y"),
  Time.name = "time",
  background = NULL,
  bg.axes = TRUE,
  bg.misc = NULL,
  bg.opts = NULL,
  blur.size = 8,
  covariate = NULL,
  covariate.colors = c("black", "white"),
  covariate.legend.loc = "bottomright",
  covariate.thresh = NULL,
  crawl.mu.color = "black",
  crawl.plot.type = "point.tail",
  date.col = "black",
  delta.t = NULL,
  dev.opts = list(),
  dimmed = NULL,
  ID.name = NULL,
  interpolation_type = "gam",
  interval = 1/12,
  legend.loc = "topright",
  main = NULL,
  max_refit_attempts = 10,
  method = "html",
  n.frames = NULL,
  network = NULL,
  network.colors = NULL,
  network.thresh = 0.5,
  network.times = NULL,
  network.ring.trans = 1,
  network.ring.wt = 3,
  network.segment.trans = 0.5,
  network.segment.wt = 3,
  override = FALSE,
  par.opts = list(),
  paths.proj = "+proj=longlat",
  paths.tranform.crs = "+proj=aea",
  plot.date = TRUE,
  pt.alpha = 0.4,
  pt.cex = 1,
  pt.colors = NULL,
  pt.wd = 1,
  res = 1.5,
  return.paths = FALSE,
  s_args = NULL,
  simulation = FALSE,
  simulation.iter = 12,
  tail.alpha = 0.6,
  tail.colors = "gray87",
  tail.length = 5,
  tail.wd = 1,
  theme_map = NULL,
  times = NULL,
  uncertainty.level = NA,
  uncertainty.type = 1,
  whole.path = FALSE,
  xlim = NULL,
  ylim = NULL,
  verbose = FALSE,
  ...
)

Arguments

paths

Either a data.frame with longitudes/eastings, latitudes/northings, IDs, and times (see coord, ID.name, and Time.name), a SpatialPointsDataFrame with IDs and times, or a list of data.frames containing the longitudes, latitudes, and times for each individual (with names provided). If all paths are already synchronous, another option for passing the data is to define paths as a list of matrices, all with the same number of rows, and to specify the times separately via the next argument. This situation might arise when, for example, locations the user wishes to animated correspond to realizations/sampler from a discrete-time movement model. Covariates may be provided as named columns of the matrices in paths.

coord

A character vector of length 2 giving the names of the longitude/easting and latitude/northing columns in the paths data.frame (in that order). This is required if paths is not a SpatialPointsDataFrame.

Time.name

The name of the columns in paths gving the observation times. This column must be of class POSIXt, or numeric.

background

Three possibilities: (1) A single background image over which animation will be overlayed, or a SpatRaster objects with one layers corresponding to each frame. (2) A list with values center (long/lat), zoom, and maptype (see ggmap::get_googlemap()) which will be used to generate a background for the animation based on Google maps tiles. Additional arguments may be added which will be passed to ggmap::get_googlemap(). (3) A logical value of TRUE, which will cue the function to get the best Google Map tile combination it can come up with. Note: ggmap must be installed for (2) and (3). Note: if you are calling animate_paths() several times in a short period of time you may get an error from Google for trying to pull tiles too often (e.g., Error in download.file(url, destfile = tmp, quiet = !messaging, mode = "wb") : cannot open URL 'http://maps.googleapis...'). Waiting a minute or so usually solves this.

bg.axes

logical: should animation place axis labels when using a background image (default is TRUE). If RGoogleMaps is used to produce background, labels will be "northing" and "easting". Otherwise, the strings given to coord will be used.

bg.misc

Character string which will be executed as R code after generating the background, and before adding trajectories, etc.

bg.opts

Options passed to plot() function call that makes background in each frame. For example, this could be used to specify blue ocean and gray landcover if background is a MULTIPOLYGON simple features object and bg.opts = list(bg = "dodgerblue4", col = "gray", border = "gray").

blur.size

a integer of the size for blur points; default is 8

covariate

The name of the column in paths that identifies the covariate to be mapped to a ring of color around each point.

covariate.colors

vector of colors which will be used in their given order to make a color ramp (see colorRamp())

covariate.legend.loc

either the location of the covariate legend, or NA if no legend is desired

covariate.thresh

if changed from its default value of NULL, the interpolated value of the covariate will be binarized based on this numeric value.

crawl.mu.color

color for the main predictions for crawl interpolation; default is black

crawl.plot.type

a character string of what type of the plot you wish to generate when interpolation_type = "crawl". Default is "point.tail" for points with tails; input "point" for point plot and input "blur" for blur point plot; ; input "blur.point" for blur point with tails.

date.col

default is "black"

delta.t

The gap in time between each frame in the animation. Specify one of delta.t or n.frames. If both are specified, delta.t is used.

dev.opts

Options passed to png() before creating each frame.

dimmed

Numeric vector of individuals to "dim" in the animation. Order corresponds to the order of the ID.name variable, or order of paths list.

ID.name

The name of the column in paths that identifies each individual. If left as NULL (default), a single individual is assumed.

interpolation_type

a character string of the type of interpolation. Default is "gam" for a generalized addictive model. Use "crawl" to interpolate using crawl package. Note: due to the ongoing shift in PROJ4/6 standards, warning about CRS comments may appear.

interval

Seconds per frame in animation. Default is 1/12 (or 12 frames per second).

legend.loc

passed to first argument of legend() function. Default is "topright". NA removes legend.

main

Title for each frame.

max_refit_attempts

an integer of number of resampling when the fit for crawl failed to run; default is 10

method

either "html" (default) or "mp4". The latter requires the user has installed ffmpeg (see ?animation::saveVideo()).

n.frames

The number of frames used to animate the complete time domain of the data.

network

Array of dimensions (# individuals, # individuals, n.frames) that gives a dyanmic network structure among the individuals.

network.colors

A symmetric matrix of dimension length(paths) \times length(paths) giving the colors associated with each pairwise relationship.

network.thresh

Network structure is summarized in the animation in a binary way, regardless of whether or not the network is continuously weighted or not. The value of network.thresh determines the level below which no connection is shown, and above which an active connection is shown via colored rings and connecting segments.

network.times

Numeric vector. If network time grid doesn't match n.frames, supply the times at which the network has been evaluated so it can be interpolated using smoothing splines.

network.ring.trans

transparency of network segments (default is 1)

network.ring.wt

thickness of network rings (default is 3)

network.segment.trans

transparency of network segments (default is 0.5)

network.segment.wt

thickness of network segments (default is 3)

override

Logical variable toggling where or not to override warnings about how long the animation procedure will take.

par.opts

Options passed to par() before creating each frame.

paths.proj

PROJ.4 string corresponding to the projection of the data. Default is "+proj=longlat".

paths.tranform.crs

a PROJ.4 string of coordinate projection transformation based on the animals' location; default is "+proj=aea +lat_1=30 +lat_2=70".

plot.date

Logical variable toggling date text at the time center of the animation.

pt.alpha

alpha value for the points

pt.cex

A numeric value giving the character expansion (size) of the points for each individual. Default is 1.

pt.colors

A vector of colors to be used for each individual in the animation. Default values come from Color Brewer palettes. When a network is provided, this is ignored and individuals are all colored black. If NA, no plot colors are chosen to distinguish individuals. This can be useful when making animations involving a covariate. Consider also setting legend.loc to NA in this case.

pt.wd

size of the points; default is 1

res

Resolution of images in animation. Increase this for higher quality (and larger) images.

return.paths

logical. Default is FALSE, but if TRUE then the interpolated paths are returned and no animation is produced.

s_args

Default is NULL, in which case anipaths attempts to select a reasonable number of knots for the GAM interpolation. Alternatively, the user can provide a list of arguments to mgcv::s() the same length and order as number of unique individuals (i.e., unique(paths[, ID.name])). Each entry in the list should be a named list/vector (e.g., s_args = list(list(k = 10), list(k = 12), ...)).

simulation

logical. Generate simulation predictions to have multiple projects for the animal paths; default is FALSE.

simulation.iter

an integer of how many paths the crawl model will generate; default is 5.

tail.alpha

alpha value for the tails

tail.colors

default is "gray87". Can be single color or vector of colors.

tail.length

Length of the tail trailing each individual.

tail.wd

Thickness of tail trailing behind each individual. Default is 1.

theme_map

plot theme for ggplot, default is NULL

times

If all paths are already synchronous, another option for passing the data is to define paths as a list of matrices, all with the same number of rows, and to specify the times separately via this argument.

uncertainty.level

value in (0, 1) corresponding to level at which to draw uncertainty ellipses. NA (default) results in no ellipses.

uncertainty.type

State what type of uncertainty plot 1 is default for tails more than 1 is amount of predicted trajectories for each unique individual and blurs for blur plot

whole.path

logical. If TRUE (default = FALSE), the complete interpolated trajectories will be plotted in the background of the animation. If whole.path = TRUE, consider also setting tail.length = 0.

xlim

Boundaries for plotting. If left undefined, the range of the data will be used.

ylim

Boundaries for plotting. If left undefined, the range of the data will be used.

verbose

logical; TRUE prints messages about fitting details

...

other arguments to be passed to ani.options to animation options such as the time interval between image frames.

Value

video file, possibly a directory containing the individual images, or interpolated paths.

Examples

##
vultures$POSIX <- as.POSIXct(vultures$timestamp, tz = "UTC")
vultures_paths <- vultures[vultures$POSIX > as.POSIXct("2009-03-01", origin = "1970-01-01") &
  vultures$POSIX < as.POSIXct("2009-05-01", origin = "1970-01-01"), ]
animate_paths(
  paths = vultures_paths,
  delta.t = "week",
  coord = c("location.long", "location.lat"),
  Time.name = "POSIX",
  ID.name = "individual.local.identifier"
)
## Not run: 
background <- list(
  center = c(-90, 10),
  zoom = 3,
  maptype = "satellite"
)
library(ggmap)
library(RColorBrewer)
COVARIATE <- cos(as.numeric(vultures_paths$timestamp) /
  diff(range(as.numeric(vultures_paths$timestamp))) * 4 * pi)
animate_paths(
  paths = cbind(vultures_paths, COVARIATE),
  delta.t = "week",
  coord = c("location.long", "location.lat"),
  Time.name = "POSIX", covariate = "COVARIATE",
  covariate.colors = brewer.pal(n = 9, "RdYlGn"),
  ID.name = "individual.local.identifier",
  background = background
)

# animation using crawl interpolation
animate_paths(
  paths = vultures_paths,
  delta.t = "week",
  coord = c("location.long", "location.lat"),
  Time.name = "POSIX",
  ID.name = "individual.local.identifier",
  interpolation_type = "crawl"
)

## End(Not run)

# Run to remove files generated by this function
system("rm -r js; rm -r css; rm -r images; rm index.html")