,
Benedikt Graeler [ctb],
Tom Gottfried [ctb],
Robert J. Hijmans [ctb]| Title: | Classes and Methods for Spatio-Temporal Data |
|---|---|
| Description: | Classes and methods for spatio-temporal data, including space-time regular lattices, sparse lattices, irregular data, and trajectories; utility functions for plotting data as map sequences (lattice or animation) or multiple time series; methods for spatial and temporal selection and subsetting, as well as for spatial/temporal/spatio-temporal matching or aggregation, retrieving coordinates, print, summary, etc. |
| Authors: | Edzer Pebesma [aut, cre] ,
Benedikt Graeler [ctb],
Tom Gottfried [ctb],
Robert J. Hijmans [ctb] |
| Maintainer: | Edzer Pebesma <[email protected]> |
| License: | GPL (>= 2) |
| Version: | 1.3-2 |
| Built: | 2024-11-03 05:23:15 UTC |
| Source: | https://github.com/edzer/spacetime |
Air quality data obtained from the airBase European air quality data base. Daily averages for rural background stations in Germany, 1998-2009. In addition, NUTS1 regions (states, or Bundeslaender) for Germany to illustrate spatial aggregation over irregular regions.
data(air)data(air)
see vignette on overlay and spatio-temporal aggregation in this package; the vignette on using google charts shows where the ISO_3166_2_DE table comes from.
air quality data compiled for R by Benedict Graeler; NUTS1 level data obtained from https://www.gadm.org/ .
https://www.eionet.europa.eu/etcs/etc-acm/databases/airbase
data(air) rural = STFDF(stations, dates, data.frame(PM10 = as.vector(air))) # how DE was created from DE_NUTS1: #if (require(rgeos)) # DE = gUnionCascaded(DE_NUTS1) #data(air) rural = STFDF(stations, dates, data.frame(PM10 = as.vector(air))) # how DE was created from DE_NUTS1: #if (require(rgeos)) # DE = gUnionCascaded(DE_NUTS1) #
find default time interval end points when intervals are regular
delta(x)delta(x)
x |
object of class |
;
to find the interval size for the last observation (which has no next
observation), x needs to be at least of length 2.
sequence of POSIXct time stamps, indicating the end
of the time interval, given by the next observation in x.
The last interval gets the same width of the one-but-last interval.
Edzer Pebesma
https://www.jstatsoft.org/v51/i07/
x = as.POSIXct("2000-01-01") + (0:9) * 3600 delta(x)x = as.POSIXct("2000-01-01") + (0:9) * 3600 delta(x)
Compute spatial or temporal empirical orthogonal function (EOF)
eof(x, how = c("spatial", "temporal"), returnEOFs = TRUE, ...) EOF(x, how = c("spatial", "temporal"), returnPredictions = TRUE, ...)eof(x, how = c("spatial", "temporal"), returnEOFs = TRUE, ...) EOF(x, how = c("spatial", "temporal"), returnPredictions = TRUE, ...)
x |
object of class |
how |
character; choose |
returnEOFs |
logical; if TRUE, the eigenvectors (EOFs) are returned in the form of a Spatial or xts object; if FALSE, the object returned by prcomp is returned, which can be printed, or from which a summary can be computed; see examples. |
returnPredictions |
logical; if TRUE, the functions are returned (i.e., predicted principle components, or PC scores); if FALSE, the object returned by prcomp is returned, which can be printed, or from which a summary can be computed; see examples (deprecated, see below). |
... |
arguments passed on to function prcomp; note that
|
In spatial mode, the appropriate Spatial* object. In temporal
mode, an object of class xts.
EOF is deprecated: it mixes up spatial and temporal EOFs,
and returns projections (PC scores) instead of EOFs (eigenvectors);
to compute EOFs, use eof.
if (require(gstat)) { data(wind) library(sp) wind.loc$y = as.numeric(char2dms(as.character(wind.loc[["Latitude"]]))) wind.loc$x = as.numeric(char2dms(as.character(wind.loc[["Longitude"]]))) coordinates(wind.loc) = ~x+y proj4string(wind.loc) = "+proj=longlat +datum=WGS84" # match station order to names in wide table: stations = 4:15 wind.loc = wind.loc[match(names(wind[stations]), wind.loc$Code),] row.names(wind.loc) = wind.loc$Station wind$time = ISOdate(wind$year+1900, wind$month, wind$day, 0) space = list(values = names(wind)[stations]) wind.st = stConstruct(wind[stations], space, wind$time, SpatialObj = wind.loc) # select firt 500 time steps, to limit run time: wind.st = wind.st[,1:500] wind.eof.1 = eof(wind.st) wind.eof.2 = eof(wind.st, "temporal") wind.eof.1.PCs = eof(wind.st, returnEOFs = FALSE) eof(wind.st, "temporal", returnEOFs = FALSE) summary(eof(wind.st, returnEOFs = FALSE)) summary(eof(wind.st, "temporal", returnEOFs = FALSE)) plot(eof(wind.st, "temporal", returnEOFs = FALSE)) }if (require(gstat)) { data(wind) library(sp) wind.loc$y = as.numeric(char2dms(as.character(wind.loc[["Latitude"]]))) wind.loc$x = as.numeric(char2dms(as.character(wind.loc[["Longitude"]]))) coordinates(wind.loc) = ~x+y proj4string(wind.loc) = "+proj=longlat +datum=WGS84" # match station order to names in wide table: stations = 4:15 wind.loc = wind.loc[match(names(wind[stations]), wind.loc$Code),] row.names(wind.loc) = wind.loc$Station wind$time = ISOdate(wind$year+1900, wind$month, wind$day, 0) space = list(values = names(wind)[stations]) wind.st = stConstruct(wind[stations], space, wind$time, SpatialObj = wind.loc) # select firt 500 time steps, to limit run time: wind.st = wind.st[,1:500] wind.eof.1 = eof(wind.st) wind.eof.2 = eof(wind.st, "temporal") wind.eof.1.PCs = eof(wind.st, returnEOFs = FALSE) eof(wind.st, "temporal", returnEOFs = FALSE) summary(eof(wind.st, returnEOFs = FALSE)) summary(eof(wind.st, "temporal", returnEOFs = FALSE)) plot(eof(wind.st, "temporal", returnEOFs = FALSE)) }
Wildfire occurrences in Northern Los Angeles County, California
between 1976 and 2000. The spatial units are in scaled feet, taken
from the NAD 83 state-plane coordinate system. One unit is equivalent
to 100,000 feet or 18.9 miles. The times for the points were produced
by the date package and represent the number of days since
January 1, 1960.
data(fires)data(fires)
A data frame with 313 observations with day of occurrence, x and y coordinates.
Roger Peng, taken from (non-CRAN) package ptproc,
https://www.biostat.jhsph.edu/~rpeng/software/index.html;
example code by Roger Bivand.
data(fires) fires$X <- fires$X*100000 fires$Y <- fires$Y*100000 library(sp) coordinates(fires) <- c("X", "Y") proj4string(fires) <- CRS("+init=epsg:2229 +ellps=GRS80") dates <- as.Date("1960-01-01")+(fires$Time-1) Fires <- STIDF(as(fires, "SpatialPoints"), dates, data.frame(time=fires$Time)) library(mapdata) if (require(sf)) { m <- map("county", "california", xlim=c(-119.1, -117.5), ylim=c(33.7, 35.0), plot=FALSE, fill=TRUE) m.sf <- st_transform(st_as_sfc(m), "EPSG:2229") cc <- as(m.sf, "Spatial") plot(cc, xlim=c(6300000, 6670000), ylim=c(1740000, 2120000)) plot(slot(Fires, "sp"), pch=3, add=TRUE) stplot(Fires, sp.layout=list("sp.lines", cc)) }data(fires) fires$X <- fires$X*100000 fires$Y <- fires$Y*100000 library(sp) coordinates(fires) <- c("X", "Y") proj4string(fires) <- CRS("+init=epsg:2229 +ellps=GRS80") dates <- as.Date("1960-01-01")+(fires$Time-1) Fires <- STIDF(as(fires, "SpatialPoints"), dates, data.frame(time=fires$Time)) library(mapdata) if (require(sf)) { m <- map("county", "california", xlim=c(-119.1, -117.5), ylim=c(33.7, 35.0), plot=FALSE, fill=TRUE) m.sf <- st_transform(st_as_sfc(m), "EPSG:2229") cc <- as(m.sf, "Spatial") plot(cc, xlim=c(6300000, 6670000), ylim=c(1740000, 2120000)) plot(slot(Fires, "sp"), pch=3, add=TRUE) stplot(Fires, sp.layout=list("sp.lines", cc)) }
Compute mnf from spatial, temporal, or spatio-temporal data
mnf(x, ...) ## S3 method for class 'matrix' mnf(x, ..., Sigma.Noise, use = "complete.obs") ## S3 method for class 'mts' mnf(x, ..., use = "complete.obs") ## S3 method for class 'zoo' mnf(x, ..., use = "complete.obs") ## S3 method for class 'SpatialPixelsDataFrame' mnf(x, ..., use = "complete.obs") ## S3 method for class 'SpatialGridDataFrame' mnf(x, ..., Sigma.Noise, use = "complete.obs") ## S3 method for class 'RasterStack' mnf(x, ..., use = "complete.obs") ## S3 method for class 'RasterBrick' mnf(x, ..., use = "complete.obs") ## S3 method for class 'STSDF' mnf(x, ..., use = "complete.obs", mode = "temporal") ## S3 method for class 'STFDF' mnf(x, ..., use = "complete.obs", mode = "temporal")mnf(x, ...) ## S3 method for class 'matrix' mnf(x, ..., Sigma.Noise, use = "complete.obs") ## S3 method for class 'mts' mnf(x, ..., use = "complete.obs") ## S3 method for class 'zoo' mnf(x, ..., use = "complete.obs") ## S3 method for class 'SpatialPixelsDataFrame' mnf(x, ..., use = "complete.obs") ## S3 method for class 'SpatialGridDataFrame' mnf(x, ..., Sigma.Noise, use = "complete.obs") ## S3 method for class 'RasterStack' mnf(x, ..., use = "complete.obs") ## S3 method for class 'RasterBrick' mnf(x, ..., use = "complete.obs") ## S3 method for class 'STSDF' mnf(x, ..., use = "complete.obs", mode = "temporal") ## S3 method for class 'STFDF' mnf(x, ..., use = "complete.obs", mode = "temporal")
x |
object for which an mnf method is available |
... |
ignored |
Sigma.Noise |
Noise covariance matrix; when missing, estimated from the data by using the covariance of lag-one spatial or temporal differences (MAF) |
use |
method to deal with missing values when computing covariances; see cov |
mode |
for |
Uses MAF (Min/max Autocorrelation Factors) to estimate the
noise covariance. This implementation estimates the noise covariance
by , so that eigenvalues can be
directly interpreted as approximate estimates of the noice covariance.
object of class (c("mnf", "prcomp"); see prcomp. Additional elements
are values, containing the eigenvalues.
https://r-spatial.org/r/2016/03/09/MNF-PCA-EOF.html
# temporal data: set.seed(13531) # make reproducible s1 = arima.sim(list(ma = rep(1,20)), 500) s2 = arima.sim(list(ma = rep(1,20)), 500) s3 = arima.sim(list(ma = rep(1,20)), 500) s3 = s3 + rnorm(500, sd = 10) d = cbind(s1,s2,s3) plot(d) m = mnf(d) m summary(m) plot(predict(m)) # spatial example: ## Not run: library(sp) grd = SpatialPoints(expand.grid(x=1:100, y=1:100)) gridded(grd) = TRUE fullgrid(grd) = TRUE pts = spsample(grd, 50, "random") pts$z = rnorm(50) library(gstat) v = vgm(1, "Sph", 90) out = krige(z~1, pts, grd, v, nmax = 20, nsim = 4) out[[3]] = 0.5 * out[[3]] + 0.5 * rnorm(1e4) out[[4]] = rnorm(1e4) spplot(out, as.table = TRUE) m = mnf(out) m summary(m) ## End(Not run) if (require(gstat)) { data(wind) library(sp) wind.loc$y = as.numeric(char2dms(as.character(wind.loc[["Latitude"]]))) wind.loc$x = as.numeric(char2dms(as.character(wind.loc[["Longitude"]]))) coordinates(wind.loc) = ~x+y proj4string(wind.loc) = "+proj=longlat +datum=WGS84" # match station order to names in wide table: stations = 4:15 wind.loc = wind.loc[match(names(wind[stations]), wind.loc$Code),] row.names(wind.loc) = wind.loc$Station wind$time = ISOdate(wind$year+1900, wind$month, wind$day, 0) space = list(values = names(wind)[stations]) wind.st = stConstruct(wind[stations], space, wind$time, SpatialObj = wind.loc, interval = TRUE) m = mnf(wind.st) m plot(m) stplot(predict(m), mode = "tp") }# temporal data: set.seed(13531) # make reproducible s1 = arima.sim(list(ma = rep(1,20)), 500) s2 = arima.sim(list(ma = rep(1,20)), 500) s3 = arima.sim(list(ma = rep(1,20)), 500) s3 = s3 + rnorm(500, sd = 10) d = cbind(s1,s2,s3) plot(d) m = mnf(d) m summary(m) plot(predict(m)) # spatial example: ## Not run: library(sp) grd = SpatialPoints(expand.grid(x=1:100, y=1:100)) gridded(grd) = TRUE fullgrid(grd) = TRUE pts = spsample(grd, 50, "random") pts$z = rnorm(50) library(gstat) v = vgm(1, "Sph", 90) out = krige(z~1, pts, grd, v, nmax = 20, nsim = 4) out[[3]] = 0.5 * out[[3]] + 0.5 * rnorm(1e4) out[[4]] = rnorm(1e4) spplot(out, as.table = TRUE) m = mnf(out) m summary(m) ## End(Not run) if (require(gstat)) { data(wind) library(sp) wind.loc$y = as.numeric(char2dms(as.character(wind.loc[["Latitude"]]))) wind.loc$x = as.numeric(char2dms(as.character(wind.loc[["Longitude"]]))) coordinates(wind.loc) = ~x+y proj4string(wind.loc) = "+proj=longlat +datum=WGS84" # match station order to names in wide table: stations = 4:15 wind.loc = wind.loc[match(names(wind[stations]), wind.loc$Code),] row.names(wind.loc) = wind.loc$Station wind$time = ISOdate(wind$year+1900, wind$month, wind$day, 0) space = list(values = names(wind)[stations]) wind.st = stConstruct(wind[stations], space, wind$time, SpatialObj = wind.loc, interval = TRUE) m = mnf(wind.st) m plot(m) stplot(predict(m), mode = "tp") }
replace NA attribute values in time series, using last or next observation, or using (temporal) interpolation, and disaggregation
## S3 method for class 'STFDF' na.locf(object, na.rm = FALSE, ...) ## S3 method for class 'STFDF' na.approx(object, x = time(object), xout, ..., na.rm = TRUE) ## S3 method for class 'STFDF' na.spline(object, x = time(object), xout, ..., na.rm = TRUE)## S3 method for class 'STFDF' na.locf(object, na.rm = FALSE, ...) ## S3 method for class 'STFDF' na.approx(object, x = time(object), xout, ..., na.rm = TRUE) ## S3 method for class 'STFDF' na.spline(object, x = time(object), xout, ..., na.rm = TRUE)
object |
object of class |
na.rm |
logical; need non-replaced NA values be removed? |
x |
times at which observations are taken; should not be modified |
xout |
if present, new times at which the time series should be approximated (disaggregated) |
... |
passed on to underlying zoo functions; see details |
details are found in na.locf, na.approx, na.spline.
object of class STFDF, with NA values replaced.
Edzer Pebesma
https://www.jstatsoft.org/v51/i07/
# toy example: library(sp) pts = SpatialPoints(cbind(c(0,1),c(0,1))) Sys.setenv(TZ="GMT") tm = seq(as.POSIXct("2012-11-25"), as.POSIXct("2012-11-30"), "1 day") df = data.frame(a = c(NA,NA,2,3,NA,NA,NA,2,NA,NA,4,NA), b = c(NA,2,3,4,5,1,2,NA,NA,NA,NA,3)) x = STFDF(pts, tm, df) as(x, "xts") as(na.locf(x), "xts") as(na.locf(x, fromLast = TRUE), "xts") as(na.locf(na.locf(x), fromLast = TRUE), "xts") # drops first record: as(na.approx(x[,,1]), "xts") # keep it: cbind(as(na.approx(x[,,1], na.rm=FALSE), "xts"), as(na.approx(x[,,2]), "xts")) cbind(as(na.spline(x[,,1]), "xts"), as(na.spline(x[,,2]), "xts")) #disaggregate: xout = seq(start(x), end(x), "6 hours") as(na.approx(x[,,1], xout = xout), "xts") as(na.spline(x[,,1], xout = xout), "xts") as(na.spline(x[,,2], xout = xout), "xts") # larger/real data: data(air) rural = STFDF(stations, dates, data.frame(PM10 = as.vector(air))) # fill NA's with last non-NA r = na.locf(rural) # sample (NOT aggregate) to monthly: m = seq(start(rural), end(rural), "1 month") stplot(na.approx(rural[1:20,"2003::2005"], xout = m), mode = 'ts')# toy example: library(sp) pts = SpatialPoints(cbind(c(0,1),c(0,1))) Sys.setenv(TZ="GMT") tm = seq(as.POSIXct("2012-11-25"), as.POSIXct("2012-11-30"), "1 day") df = data.frame(a = c(NA,NA,2,3,NA,NA,NA,2,NA,NA,4,NA), b = c(NA,2,3,4,5,1,2,NA,NA,NA,NA,3)) x = STFDF(pts, tm, df) as(x, "xts") as(na.locf(x), "xts") as(na.locf(x, fromLast = TRUE), "xts") as(na.locf(na.locf(x), fromLast = TRUE), "xts") # drops first record: as(na.approx(x[,,1]), "xts") # keep it: cbind(as(na.approx(x[,,1], na.rm=FALSE), "xts"), as(na.approx(x[,,2]), "xts")) cbind(as(na.spline(x[,,1]), "xts"), as(na.spline(x[,,2]), "xts")) #disaggregate: xout = seq(start(x), end(x), "6 hours") as(na.approx(x[,,1], xout = xout), "xts") as(na.spline(x[,,1], xout = xout), "xts") as(na.spline(x[,,2], xout = xout), "xts") # larger/real data: data(air) rural = STFDF(stations, dates, data.frame(PM10 = as.vector(air))) # fill NA's with last non-NA r = na.locf(rural) # sample (NOT aggregate) to monthly: m = seq(start(rural), end(rural), "1 month") stplot(na.approx(rural[1:20,"2003::2005"], xout = m), mode = 'ts')
convert a spatial nb object to a matching STF object
nbMult(nb, st, addT = TRUE, addST = FALSE)nbMult(nb, st, addT = TRUE, addST = FALSE)
nb |
object of class nb (see package spdep), which is valid for
the spatial slot of object |
st |
object of class STF |
addT |
logical; should temporal neighbours be added? |
addST |
logical; should spatio-temporal neighbours be added? |
if both addT and addST are false, only
spatial neighbours are added for each time replicate.
details are found in
Giovana M. de Espindola, Edzer Pebesma, Gilberto Câmara, 2011. Spatio-temporal regression models for deforestation in the Brazilian Amazon. STDM 2011, The International Symposium on Spatial-Temporal Analysis and Data Mining, University College London - 18th-20th July 2011.
object of class nb
Edzer Pebesma
consistent spatio-temporal overlay for STF, STS and STI objects, as well as their *DF counterpart: retrieves the indexes or attributes from one geometry at the spatio-temporal points of another
## S4 method for signature 'STF,STF' over(x, y, returnList = FALSE, fn = NULL, ...) ## S4 method for signature 'xts,xts' over(x, y, returnList = FALSE, fn = NULL, ...) ## S4 method for signature 'ST' aggregate(x, by, FUN, ..., simplify = TRUE)## S4 method for signature 'STF,STF' over(x, y, returnList = FALSE, fn = NULL, ...) ## S4 method for signature 'xts,xts' over(x, y, returnList = FALSE, fn = NULL, ...) ## S4 method for signature 'ST' aggregate(x, by, FUN, ..., simplify = TRUE)
x |
geometry (S/T locations) of the queries |
y |
layer from which the geometries or attributes are queried |
returnList |
logical; determines whether a list is returned, or an index vector |
fn |
(optional) a function; see value |
by |
geometry over which attributes in |
FUN |
aggregation function |
simplify |
boolean; if TRUE, and space or time dimensions can be
dropped, the simpler ( |
... |
arguments passed on to function fn or FUN |
an object of length length(x), or a data.frame with number
of rows equal to length(x). If returnList is FALSE,
a vector with indices of y for each geometry (point, grid
cell centre, polygon or lines x time point) in x. if returnList is
TRUE, a list of length length(x), with list element i
the vector of indices of the geometries in y that correspond
to the $i$-th geometry in x.
The aggregate method for ST objects aggregates the attribute values of x
over the geometry (space, time, or space-time) of by, using
aggregation function FUN.
For the matching of time intervals, see timeMatch.
For setting, or retrieving whether time represents intervals, see timeIsInterval.
finds the row index of the instance
or interval of time instances of x matching to y. Only if
timeIsInterval(x) == TRUE, intervals are sought. In that case,
time intervals start at the time instance of a record, and end at
the next. The last time interval length is set to the interval length
of the one-but-last (non-zero) interval. In case of a single time
instance for y, its interval is right-open.
See also over; methods intersecting SpatialLines with anything else, or SpatialPolygons with SpatialPolygons, need rgeos to be loaded first.
Edzer Pebesma, [email protected]
https://www.jstatsoft.org/article/view/v051i07
over; vignette('sto'), vignette('over'),
timeMatch, timeIsInterval
read or write tgrass (time-enabled grass) files
read.tgrass(fname, localName = TRUE, useTempDir = TRUE, isGeoTiff = TRUE) write.tgrass(obj, fname, ...)read.tgrass(fname, localName = TRUE, useTempDir = TRUE, isGeoTiff = TRUE) write.tgrass(obj, fname, ...)
fname |
file name to read from, or write to |
localName |
logical; if TRUE, |
useTempDir |
logical: use a temporary directory for extraction? |
isGeoTiff |
logical: are the files in the tar.gz file GeoTIFFs? |
obj |
object to export, of class |
... |
arguments passed on to writeRaster |
The tgrass format is a gzip'ed tar file (.tar.gz) that has geotiff files (with suffix .tif), and three files (list.txt, proj.txt and init.txt) describing the file names and time slices, coordinate reference system, and dimensions
read.tgrass returns an object of class RasterStack,
writegrass returns nothing
Edzer Pebesma; time-enabled grass by Soeren Gebbert
https://dx.doi.org/10.1016/j.envsoft.2013.11.001
## Not run: library(spacetime) r = read.tgrass("precipitation_1950_2011_yearly.tar.gz", useTempDir = FALSE) write.tgrass(r, "myfile.tar.gz") ## End(Not run)## Not run: library(spacetime) r = read.tgrass("precipitation_1950_2011_yearly.tar.gz", useTempDir = FALSE) write.tgrass(r, "myfile.tar.gz") ## End(Not run)
An abstract class from which useful spatio-temporal classes are derived
ST(sp, time, endTime)ST(sp, time, endTime)
sp |
an object deriving from class Spatial, such as a SpatialPoints or SpatialPolygons |
time |
an object of class xts, or a time vector (currently: Date, POSIXct, timeDate, yearmon and yearqtr; are supported; see xts); in the latter case, it should be in time order |
endTime |
vector of class |
Objects of this class are not meant to be useful; only derived classes can be meaningful
sp:Object deriving from class "Spatial"
time:Object of class "xts"
signature(obj = "ST"): retrieves the attribute element
signature(obj = "ST"): retrieves the attribute element
signature(obj = "ST"): sets or replaces the attribute element
signature(obj = "ST"): sets or replaces the attribute element
argument (and object slot) sp can be pure geometry, or geometry with attributes. In the latter case, the geometries are kept with the sp slot, and only replicated (when needed) on coercion to the long format, with as.data.frame.
Slot time needs to be of class xts; if a time or date vector is passed as argument to SP, it will be converted into an xts object.
When endTime is missing, an error is thrown.
ST is meant as a super-class, and is not to be used for representing data, similar to Spatial in the sp package.
Edzer Pebesma, [email protected]
https://www.jstatsoft.org/v51/i07/
time = as.Date('2008-01-01')+1:2 library(sp) sp = SpatialPoints(cbind(c(0,1),c(0,1))) ST(sp, time, delta(time))time = as.Date('2008-01-01')+1:2 library(sp) sp = SpatialPoints(cbind(c(0,1),c(0,1))) ST(sp, time, delta(time))
obtain ranges of space and time coordinates
stbox(obj) bbox(obj)stbox(obj) bbox(obj)
obj |
object of a class deriving from |
stbox returns a data.frame, with three columns
representing x-, y- and time-coordinates, and two rows containing min
and max values. bbox gives a matrix with coordinate min/max values,
compatible to bbox
signature(x = "ST"): obtain st range from object
create ST* objects from long or wide tables
stConstruct(x, space, time, SpatialObj = NULL, TimeObj = NULL, crs = CRS(as.character(NA)), interval, endTime)stConstruct(x, space, time, SpatialObj = NULL, TimeObj = NULL, crs = CRS(as.character(NA)), interval, endTime)
x |
object of class |
space |
in case |
time |
in case |
SpatialObj |
object of class Spatial-class, containing the locations of a time-wide table, or the locations of a long table |
TimeObj |
in case of space-wide table,
object of class xts, containing the times for each
of the columns in a list element of |
crs |
object of class CRS-class; only used when coordinates
are in |
interval |
logical; specifies whether time should reflect time instance (FALSE) or time intervals (TRUE). If omitted, defaults values depend on the class |
endTime |
vector of |
For examples, see below.
A long table is a data.frame with each row holding a single observation in space-time, and particular columns in this table indicate the space (location or location ID) and time.
A space-wide table is a table in which different columns refer to different locations, and each row reflects a particular observation time.
A time-wide table is a table where different times of a particular characteristic are represented as different colunns; rows in the table represent particular locations or location IDs.
Depending on the arguments, an object of class STIDF
or STFDF.
https://www.jstatsoft.org/v51/i07/
# stConstruct multivariable, time-wide if (require(maps) && require(plm) && require(sf)) { library(sp) states.m <- map('state', plot=FALSE, fill=TRUE) IDs <- sapply(strsplit(states.m$names, ":"), function(x) x[1]) sf = st_as_sf(states.m, IDs=IDs) row.names(sf) = sf$ID # not needed if sf >= 1.0-13 states <- as(sf, "Spatial") states=geometry(states) yrs = 1970:1986 time = as.POSIXct(paste(yrs, "-01-01", sep=""), tz = "GMT") data("Produc") # deselect District of Columbia, polygon 8, which is not present in Produc: Produc.st <- STFDF(states[-8], time, Produc[order(Produc[,2], Produc[,1]),]) # stplot(Produc.st[,,"unemp"], yrs, col.regions = brewer.pal(9, "YlOrRd"),cuts=9) # example 1: st from long table, with states as Spatial object: # use Date format for time: Produc$time = as.Date(paste(yrs, "01", "01", sep = "-")) # take centroids of states: xy = coordinates(states[-8]) Produc$x = xy[,1] Produc$y = xy[,2] #using stConstruct, use polygon centroids for location: x = stConstruct(Produc, c("x", "y"), "time", interval = TRUE) class(x) stplot(x[,,"unemp"]) # alternatively, pass states as SpatialObj: Produc$state = gsub("TENNESSE", "TENNESSEE", Produc$state) Produc$State = gsub("_", " ", tolower(Produc$state)) x = stConstruct(Produc, "State", "time", states) class(x) all.equal(x, Produc.st, check.attributes = FALSE) } if (require(sf)) { fname = system.file("shape/nc.shp", package="sf")[1] nc = as(st_read(fname), "Spatial") timesList = list( BIR=c("BIR74", "BIR79"), # sets of variables that belong together NWBIR=c("NWBIR74", "NWBIR79"), # only separated by space SID=c("SID74", "SID79") ) t = as.Date(c("1974-01-01","1979-01-01")) nc.st = stConstruct(as(nc, "data.frame"), geometry(nc), timesList, TimeObj = t, interval = TRUE) } # stConstruct multivariable, space-wide if (require(gstat)) { data(wind) wind.loc$y = as.numeric(char2dms(as.character(wind.loc[["Latitude"]]))) wind.loc$x = as.numeric(char2dms(as.character(wind.loc[["Longitude"]]))) coordinates(wind.loc) = ~x+y proj4string(wind.loc) = "+proj=longlat +datum=WGS84" # match station order to names in wide table: stations = 4:15 wind.loc = wind.loc[match(names(wind[stations]), wind.loc$Code),] row.names(wind.loc) = wind.loc$Station # convert to utm zone 29, to be able to do interpolation in # proper Euclidian (projected) space: # create time variable wind$time = ISOdate(wind$year+1900, wind$month, wind$day, 0) w = STFDF(wind.loc, wind$time, data.frame(values = as.vector(t(wind[stations])))) space = list(values = names(wind)[stations]) wind.st = stConstruct(wind[stations], space, wind$time, SpatialObj = wind.loc, interval = TRUE) all.equal(w, wind.st) class(wind.st) }# stConstruct multivariable, time-wide if (require(maps) && require(plm) && require(sf)) { library(sp) states.m <- map('state', plot=FALSE, fill=TRUE) IDs <- sapply(strsplit(states.m$names, ":"), function(x) x[1]) sf = st_as_sf(states.m, IDs=IDs) row.names(sf) = sf$ID # not needed if sf >= 1.0-13 states <- as(sf, "Spatial") states=geometry(states) yrs = 1970:1986 time = as.POSIXct(paste(yrs, "-01-01", sep=""), tz = "GMT") data("Produc") # deselect District of Columbia, polygon 8, which is not present in Produc: Produc.st <- STFDF(states[-8], time, Produc[order(Produc[,2], Produc[,1]),]) # stplot(Produc.st[,,"unemp"], yrs, col.regions = brewer.pal(9, "YlOrRd"),cuts=9) # example 1: st from long table, with states as Spatial object: # use Date format for time: Produc$time = as.Date(paste(yrs, "01", "01", sep = "-")) # take centroids of states: xy = coordinates(states[-8]) Produc$x = xy[,1] Produc$y = xy[,2] #using stConstruct, use polygon centroids for location: x = stConstruct(Produc, c("x", "y"), "time", interval = TRUE) class(x) stplot(x[,,"unemp"]) # alternatively, pass states as SpatialObj: Produc$state = gsub("TENNESSE", "TENNESSEE", Produc$state) Produc$State = gsub("_", " ", tolower(Produc$state)) x = stConstruct(Produc, "State", "time", states) class(x) all.equal(x, Produc.st, check.attributes = FALSE) } if (require(sf)) { fname = system.file("shape/nc.shp", package="sf")[1] nc = as(st_read(fname), "Spatial") timesList = list( BIR=c("BIR74", "BIR79"), # sets of variables that belong together NWBIR=c("NWBIR74", "NWBIR79"), # only separated by space SID=c("SID74", "SID79") ) t = as.Date(c("1974-01-01","1979-01-01")) nc.st = stConstruct(as(nc, "data.frame"), geometry(nc), timesList, TimeObj = t, interval = TRUE) } # stConstruct multivariable, space-wide if (require(gstat)) { data(wind) wind.loc$y = as.numeric(char2dms(as.character(wind.loc[["Latitude"]]))) wind.loc$x = as.numeric(char2dms(as.character(wind.loc[["Longitude"]]))) coordinates(wind.loc) = ~x+y proj4string(wind.loc) = "+proj=longlat +datum=WGS84" # match station order to names in wide table: stations = 4:15 wind.loc = wind.loc[match(names(wind[stations]), wind.loc$Code),] row.names(wind.loc) = wind.loc$Station # convert to utm zone 29, to be able to do interpolation in # proper Euclidian (projected) space: # create time variable wind$time = ISOdate(wind$year+1900, wind$month, wind$day, 0) w = STFDF(wind.loc, wind$time, data.frame(values = as.vector(t(wind[stations])))) space = list(values = names(wind)[stations]) wind.st = stConstruct(wind[stations], space, wind$time, SpatialObj = wind.loc, interval = TRUE) all.equal(w, wind.st) class(wind.st) }
A class for spatio-temporal data with full space-time grid; for n spatial locations and m times, n x m observations are available
STF(sp, time, endTime = delta(time)) STFDF(sp, time, data, endTime = delta(time)) ## S4 method for signature 'STFDF' x[i, j, ..., drop = is(x, "STFDF")] ## S4 method for signature 'STFDF,xts' coerce(from, to, strict=TRUE) ## S4 method for signature 'STFDF,Spatial' coerce(from, to)STF(sp, time, endTime = delta(time)) STFDF(sp, time, data, endTime = delta(time)) ## S4 method for signature 'STFDF' x[i, j, ..., drop = is(x, "STFDF")] ## S4 method for signature 'STFDF,xts' coerce(from, to, strict=TRUE) ## S4 method for signature 'STFDF,Spatial' coerce(from, to)
sp |
object of class Spatial, having |
time |
object holding time information, of length |
endTime |
vector of class |
data |
data frame with |
x |
an object of class STFDF |
i |
selection of spatial entities |
j |
selection of temporal entities (see syntax in package xts) |
... |
selection of attribute(s) |
drop |
if TRUE and a single spatial entity is selected, an object
of class xts is returned; if TRUE and a single temporal entity is
selected, and object of the appropriate |
from |
object of class STFDF |
to |
target class |
strict |
ignored |
the as.data.frame coercion returns the full long table,
with purely spatial attributes and purely time attributes replicated
appropriately.
Objects of this class represent full space/time data with a full grid (or lattice) layout
sp:spatial object; see ST-class
time:temporal object; see ST-class
data:Object of class data.frame, which holds
the measured values; space index cycling first, time order preserved
signature(x = "STFDF"): selects spatial entities, temporal entities, and attributes
STFDF,xts
STFDF,Spatial
(from) coerces to (wide form) SpatialXxDataFrame, where SpatialXx is the spatial class of from@sp
signature(x = "STF", y = "missing"): plots space-time
layout
signature(x = "STFDF", y = "missing"): plots space-time
layout, indicating full missing valued records
Edzer Pebesma, [email protected]
https://www.jstatsoft.org/v51/i07/
sp = cbind(x = c(0,0,1), y = c(0,1,1)) row.names(sp) = paste("point", 1:nrow(sp), sep="") library(sp) sp = SpatialPoints(sp) time = as.POSIXct("2010-08-05")+3600*(10:13) m = c(10,20,30) # means for each of the 3 point locations mydata = rnorm(length(sp)*length(time),mean=rep(m, 4)) IDs = paste("ID",1:length(mydata)) mydata = data.frame(values = signif(mydata,3), ID=IDs) stfdf = STFDF(sp, time, mydata) stfdf stfdf[1:2,] stfdf[,1:2] stfdf[,,2] stfdf[,,"values"] stfdf[1,] stfdf[,2] as(stfdf[,,1], "xts") as(stfdf[,,2], "xts") # examples for [[, [[<-, $ and $<- stfdf[[1]] stfdf[["values"]] stfdf[["newVal"]] <- rnorm(12) stfdf$ID stfdf$ID = paste("OldIDs", 1:12, sep="") stfdf$NewID = paste("NewIDs", 12:1, sep="") stfdf x = stfdf[stfdf[1:2,],] all.equal(x, stfdf[1:2,]) all.equal(stfdf, stfdf[stfdf,]) # converts character to factor...sp = cbind(x = c(0,0,1), y = c(0,1,1)) row.names(sp) = paste("point", 1:nrow(sp), sep="") library(sp) sp = SpatialPoints(sp) time = as.POSIXct("2010-08-05")+3600*(10:13) m = c(10,20,30) # means for each of the 3 point locations mydata = rnorm(length(sp)*length(time),mean=rep(m, 4)) IDs = paste("ID",1:length(mydata)) mydata = data.frame(values = signif(mydata,3), ID=IDs) stfdf = STFDF(sp, time, mydata) stfdf stfdf[1:2,] stfdf[,1:2] stfdf[,,2] stfdf[,,"values"] stfdf[1,] stfdf[,2] as(stfdf[,,1], "xts") as(stfdf[,,2], "xts") # examples for [[, [[<-, $ and $<- stfdf[[1]] stfdf[["values"]] stfdf[["newVal"]] <- rnorm(12) stfdf$ID stfdf$ID = paste("OldIDs", 1:12, sep="") stfdf$NewID = paste("NewIDs", 12:1, sep="") stfdf x = stfdf[stfdf[1:2,],] all.equal(x, stfdf[1:2,]) all.equal(stfdf, stfdf[stfdf,]) # converts character to factor...
A class for unstructured spatio-temporal data; for n spatial locations and times, n observations are available
STI(sp, time, endTime) STIDF(sp, time, data, endTime) ## S4 method for signature 'STIDF' x[i, j, ..., drop = FALSE] ## S4 method for signature 'STIDF,STSDF' coerce(from, to, strict=TRUE)STI(sp, time, endTime) STIDF(sp, time, data, endTime) ## S4 method for signature 'STIDF' x[i, j, ..., drop = FALSE] ## S4 method for signature 'STIDF,STSDF' coerce(from, to, strict=TRUE)
sp |
object of class Spatial |
time |
object holding time information; when STIDF is called,
a non-ordered vector with times, e.g. POSIXct will also work,
and rearrange the |
endTime |
vector of class |
data |
data frame with appropriate number of rows |
x |
an object of class STFDF |
i |
selection of record index (spatial/temporal/spatio-temporal entities) |
j |
or character string with temporal selection |
... |
first element is taken as column (variable) selector |
drop |
if TRUE and a single spatial entity is selected, an object
of class xts is returned (NOT yet implemented);
if TRUE and a single temporal entity is
selected, and object of the appropriate |
from |
object of class STFDF |
to |
target class |
strict |
ignored |
Objects of this class carry full space/time grid data
sp:Object of class "Spatial"
time:Object holding time information, see ST-class
data:Object of class data.frame, which holds
the measured values
signature(x = "STIDF"): selects spatial-temporal entities, and attributes
arguments sp, time and data need
to have the same number of records,
and regardless of the class of time (xts or POSIXct) have to be
in correspoinding order: the triple sp[i], time[i]
and data[i,] refer to the same observation
Edzer Pebesma, [email protected]
https://www.jstatsoft.org/v51/i07/
sp = cbind(x = c(0,0,1), y = c(0,1,1)) row.names(sp) = paste("point", 1:nrow(sp), sep="") library(sp) sp = SpatialPoints(sp) time = as.POSIXct("2010-08-05")+3600*(10:13) m = c(10,20,30) # means for each of the 3 point locations mydata = rnorm(length(sp)*length(time),mean=rep(m, 4)) IDs = paste("ID",1:length(mydata)) mydata = data.frame(values = signif(mydata,3), ID=IDs) stidf = as(STFDF(sp, time, mydata), "STIDF") stidf[1:2,] all.equal(stidf, stidf[stidf,])sp = cbind(x = c(0,0,1), y = c(0,1,1)) row.names(sp) = paste("point", 1:nrow(sp), sep="") library(sp) sp = SpatialPoints(sp) time = as.POSIXct("2010-08-05")+3600*(10:13) m = c(10,20,30) # means for each of the 3 point locations mydata = rnorm(length(sp)*length(time),mean=rep(m, 4)) IDs = paste("ID",1:length(mydata)) mydata = data.frame(values = signif(mydata,3), ID=IDs) stidf = as(STFDF(sp, time, mydata), "STIDF") stidf[1:2,] all.equal(stidf, stidf[stidf,])
subtract marginal (spatial and temporal) means from observations
stInteraction(x, ...)stInteraction(x, ...)
x |
object of class |
... |
arguments passed to rowMeans, colMeans and mean, such as |
object of class STFDF with each attribute replaced by its
residual, computed by with the
grand mean, the temporal mean, the spatial mean
and the grand mean.
if (require(gstat)) { library(sp) data(wind) wind.loc$y = as.numeric(char2dms(as.character(wind.loc[["Latitude"]]))) wind.loc$x = as.numeric(char2dms(as.character(wind.loc[["Longitude"]]))) coordinates(wind.loc) = ~x+y proj4string(wind.loc) = "+proj=longlat +datum=WGS84" # match station order to names in wide table: stations = 4:15 wind.loc = wind.loc[match(names(wind[stations]), wind.loc$Code),] row.names(wind.loc) = wind.loc$Station wind$time = ISOdate(wind$year+1900, wind$month, wind$day, 0) space = list(values = names(wind)[stations]) wind.st = stConstruct(wind[stations], space, wind$time, SpatialObj = wind.loc) wind.sti = stInteraction(wind.st) # temporal means for any station should be zero: c(mean(wind.sti[3,]), # spatial mean for each time step should be zero: mean(wind.sti[,5][[1]])) }if (require(gstat)) { library(sp) data(wind) wind.loc$y = as.numeric(char2dms(as.character(wind.loc[["Latitude"]]))) wind.loc$x = as.numeric(char2dms(as.character(wind.loc[["Longitude"]]))) coordinates(wind.loc) = ~x+y proj4string(wind.loc) = "+proj=longlat +datum=WGS84" # match station order to names in wide table: stations = 4:15 wind.loc = wind.loc[match(names(wind[stations]), wind.loc$Code),] row.names(wind.loc) = wind.loc$Station wind$time = ISOdate(wind$year+1900, wind$month, wind$day, 0) space = list(values = names(wind)[stations]) wind.st = stConstruct(wind[stations], space, wind$time, SpatialObj = wind.loc) wind.sti = stInteraction(wind.st) # temporal means for any station should be zero: c(mean(wind.sti[3,]), # spatial mean for each time step should be zero: mean(wind.sti[,5][[1]])) }
create trellis plot for ST objects
stplot(obj, ...) stplot.STFDF(obj, names.attr = trimDates(obj), ..., as.table = TRUE, at, cuts = 15, scales = list(draw = FALSE), animate = 0, mode = "xy", scaleX = 0, auto.key = list(space = key.space), main, key.space = "right", type = "l", do.repeat = TRUE, range.expand = 0.001) stplot.STIDF(obj, ..., names.attr = NULL, as.table = TRUE, scales = list(draw = FALSE), xlab = NULL, ylab = NULL, type = "p", number = 6, tcuts, sp.layout = NULL, xlim = bbox(obj)[1, ], ylim = bbox(obj)[2, ])stplot(obj, ...) stplot.STFDF(obj, names.attr = trimDates(obj), ..., as.table = TRUE, at, cuts = 15, scales = list(draw = FALSE), animate = 0, mode = "xy", scaleX = 0, auto.key = list(space = key.space), main, key.space = "right", type = "l", do.repeat = TRUE, range.expand = 0.001) stplot.STIDF(obj, ..., names.attr = NULL, as.table = TRUE, scales = list(draw = FALSE), xlab = NULL, ylab = NULL, type = "p", number = 6, tcuts, sp.layout = NULL, xlim = bbox(obj)[1, ], ylim = bbox(obj)[2, ])
obj |
object of a class deriving from |
names.attr |
names that will be used in the strip; trimDates(obj) trims "-01" ending(s) from printed Dates |
as.table |
logical; if TRUE, time will increas from top to bottom; if FALSE, time will increase from bottom to top |
at |
values at which colours will change; see levelplot |
cuts |
number of levels the range of the attribute would be divided into |
animate |
numeric; if larger than 0, the number of seconds between subsequent animated time steps (loop; press ctrl-C or Esc to stop) |
mode |
plotting mode; if "xy", maps for time steps are plotted;
if "xt", a space-time plot is constructed (see argument |
scaleX |
integer: 0, 1 or 2; when |
auto.key |
see the |
main |
character; plot title, use |
key.space |
character; see xyplot |
scales |
scales drawing; see |
xlab |
x-axis label |
ylab |
y-axis label |
type |
character; use 'l' for lines, 'p' for symbols, 'b' for both lines and symbols |
do.repeat |
logical; repeat the animation in an infinite loop? |
range.expand |
numeric; if |
number |
number of time intervals, equally spaced |
tcuts |
time cuts in units of |
sp.layout |
list or NULL; see spplot |
... |
arguments passed on to spplot in case of
plotting objects of class |
xlim |
numeric, x range |
ylim |
numeric, y range |
In non-animation and "xy" mode, stplot is a wrapper
around spplot, that automically plots each time stamp in
a panel. The returned value is is a lattice plot.
In "xt" mode, a space-time plot with space on the x-axis and time
on the y-axis is plotted. By default, the space ID is plotted on the
x-axis, as space can be anything (points, polygons, grid cells etc).
When scaleX is set to 1 or 2, the x- resp. y-coordinates
of the spatial locations, obtained by coordinates, is
used instead. Beware: when the x-coordinate is plotted, and for
each (x,t) element multiple y-coordinates are sent to the plot,
it is not clear which (x,y,t) value becomes the plotted value,
so slicing single y values is adviced – no checking is done.
The returned value is is a lattice plot.
In animation mode (animate > 0), single maps are animated
in an endless loop, with animate seconds between each. No
proper value is returned: the loop needs to be interrupted by
the user.
signature(x = "STFDF"): plots object of class STFDF
signature(x = "STSDF"): plots object of class STSDF
signature(x = "STI"): plots object of class STI
signature(x = "STIDF"): plots object of class STIDF
signature(x = "STT"): plots object of class STT
signature(x = "STTDF"): plots object of class STTDF
vignette("spacetime") contains several examples
https://www.jstatsoft.org/v51/i07/
A class for spatio-temporal data with partial space-time grids; for n spatial locations and m times, an index table is kept for which nodes observations are available
STS(sp, time, index, endTime = delta(time)) STSDF(sp, time, data, index, endTime = delta(time)) ## S4 method for signature 'STSDF' x[i, j, ..., drop = is(x, "STSDF")] ## S4 method for signature 'STSDF,STFDF' coerce(from, to, strict=TRUE) ## S4 method for signature 'STSDF,STIDF' coerce(from, to, strict=TRUE)STS(sp, time, index, endTime = delta(time)) STSDF(sp, time, data, index, endTime = delta(time)) ## S4 method for signature 'STSDF' x[i, j, ..., drop = is(x, "STSDF")] ## S4 method for signature 'STSDF,STFDF' coerce(from, to, strict=TRUE) ## S4 method for signature 'STSDF,STIDF' coerce(from, to, strict=TRUE)
sp |
object of class Spatial |
time |
object holding time information; see ST-class |
data |
data frame with rows corresponding to the observations (spatial index moving faster than temporal) |
index |
two-column matrix: rows corresponding to the nodes for which observations are available, first column giving spatial index, second column giving temporal index |
endTime |
vector of class |
x |
an object of class STFDF |
i |
selection of spatial entities |
j |
selection of temporal entities (see syntax in package xts) |
... |
selection of attribute(s) |
drop |
if TRUE and a single spatial entity is selected, an object
of class xts is returned; if TRUE and a single temporal entity is
selected, and object of the appropriate |
from |
object of class STFDF |
to |
target class |
strict |
ignored |
Objects of this class carry sparse space/time grid data
sp:Object of class "Spatial"
time:Object holding time information; see ST-class for permitted types
index:matrix of dimension n x 2, where n matches the number of rows in slot data
data:Object of class data.frame, which holds
the measured values
signature(x = "STSDF"): selects spatial entities, temporal entities, and attributes
signature(x = "STS", y = "missing"): plots
space-time layout
signature(x = "STSDF", y = "missing"): plots
space-time layout, indicating records partially NA
Edzer Pebesma, [email protected]
https://www.jstatsoft.org/v51/i07/
sp = cbind(x = c(0,0,1), y = c(0,1,1)) row.names(sp) = paste("point", 1:nrow(sp), sep="") library(sp) sp = SpatialPoints(sp) library(xts) time = xts(1:4, as.POSIXct("2010-08-05")+3600*(10:13)) m = c(10,20,30) # means for each of the 3 point locations mydata = rnorm(length(sp)*length(time),mean=rep(m, 4)) IDs = paste("ID",1:length(mydata)) mydata = data.frame(values = signif(mydata,3), ID=IDs) stfdf = STFDF(sp, time, mydata) stfdf stsdf = as(stfdf, "STSDF") stsdf[1:2,] stsdf[,1:2] stsdf[,,2] stsdf[,,"values"] stsdf[1,] stsdf[,2] # examples for [[, [[<-, $ and $<- stsdf[[1]] stsdf[["values"]] stsdf[["newVal"]] <- rnorm(12) stsdf$ID stsdf$ID = paste("OldIDs", 1:12, sep="") stsdf$NewID = paste("NewIDs", 12:1, sep="") stsdf x = stsdf[stsdf,] x = stsdf[stsdf[1:2,],] all.equal(x, stsdf[1:2,])sp = cbind(x = c(0,0,1), y = c(0,1,1)) row.names(sp) = paste("point", 1:nrow(sp), sep="") library(sp) sp = SpatialPoints(sp) library(xts) time = xts(1:4, as.POSIXct("2010-08-05")+3600*(10:13)) m = c(10,20,30) # means for each of the 3 point locations mydata = rnorm(length(sp)*length(time),mean=rep(m, 4)) IDs = paste("ID",1:length(mydata)) mydata = data.frame(values = signif(mydata,3), ID=IDs) stfdf = STFDF(sp, time, mydata) stfdf stsdf = as(stfdf, "STSDF") stsdf[1:2,] stsdf[,1:2] stsdf[,,2] stsdf[,,"values"] stsdf[1,] stsdf[,2] # examples for [[, [[<-, $ and $<- stsdf[[1]] stsdf[["values"]] stsdf[["newVal"]] <- rnorm(12) stsdf$ID stsdf$ID = paste("OldIDs", 1:12, sep="") stsdf$NewID = paste("NewIDs", 12:1, sep="") stsdf x = stsdf[stsdf,] x = stsdf[stsdf[1:2,],] all.equal(x, stsdf[1:2,])
A class for spatio-temporal trajectory data
## S4 method for signature 'STTDF,ltraj' coerce(from, to, strict=TRUE) ## S4 method for signature 'ltraj,STTDF' coerce(from, to, strict=TRUE)## S4 method for signature 'STTDF,ltraj' coerce(from, to, strict=TRUE) ## S4 method for signature 'ltraj,STTDF' coerce(from, to, strict=TRUE)
from |
from object |
to |
target class |
strict |
ignored |
Objects of this class carry sparse (irregular) space/time data
sp:Object of class "Spatial", containing the bounding
box of all trajectories
time:Object of class "xts", containing the temporal
bounding box of all trajectories
traj:Object of class list, each element holding
an STI object reflecting a single trajectory;
data:Object of class data.frame, which holds
the data values for each feature in each trajectory
signature(x = "STTDF"): select trajectories, based
on index, or spatial and/or temporal predicates
The data.frame needs to have a column called
burst which is a factor (or character) and contains the
grouping of observations that come from a continuous sequence of
observations. In addition, a column id is used to identify
individual items.
Edzer Pebesma, [email protected]
https://www.jstatsoft.org/v51/i07/
library(sp) m = 3# nr of trajectories n = 100 # length of each l = vector("list", m) t0 = as.POSIXct("2013-05-05",tz="GMT") set.seed(1331) # fix randomness for (i in 1:m) { x = cumsum(rnorm(n)) y = cumsum(rnorm(n)) sp = SpatialPoints(cbind(x,y)) #t = t0 + (0:(n-1) + (i-1)*n) * 60 t = t0 + (0:(n-1) + (i-1)*n/2) * 60 l[[i]] = STI(sp, t) } stt= STT(l) sttdf = STTDF(stt, data.frame(attr = rnorm(n*m), id = paste("ID", rep(1:m, each=n)))) x = as(stt, "STI") stplot(sttdf, col=1:m, scales=list(draw=TRUE)) stplot(sttdf, by = "id") stplot(sttdf[1]) stplot(sttdf[1]) # select a trajectory that intersect with a polygon p = Polygon(cbind(x=c(-20,-15,-15,-20,-20),y=c(10,10,15,15,10))) pol=SpatialPolygons(list(Polygons(list(p), "ID"))) #if (require(rgeos)) { # stplot(sttdf[pol]) # names(sttdf[pol]@traj) # stplot(sttdf[1:2],col=1:2) # stplot(sttdf[,t0]) # stplot(sttdf[,"2013"]) # stplot(sttdf[pol,"2013"]) # is.null(sttdf[pol,t0]) #}library(sp) m = 3# nr of trajectories n = 100 # length of each l = vector("list", m) t0 = as.POSIXct("2013-05-05",tz="GMT") set.seed(1331) # fix randomness for (i in 1:m) { x = cumsum(rnorm(n)) y = cumsum(rnorm(n)) sp = SpatialPoints(cbind(x,y)) #t = t0 + (0:(n-1) + (i-1)*n) * 60 t = t0 + (0:(n-1) + (i-1)*n/2) * 60 l[[i]] = STI(sp, t) } stt= STT(l) sttdf = STTDF(stt, data.frame(attr = rnorm(n*m), id = paste("ID", rep(1:m, each=n)))) x = as(stt, "STI") stplot(sttdf, col=1:m, scales=list(draw=TRUE)) stplot(sttdf, by = "id") stplot(sttdf[1]) stplot(sttdf[1]) # select a trajectory that intersect with a polygon p = Polygon(cbind(x=c(-20,-15,-15,-20,-20),y=c(10,10,15,15,10))) pol=SpatialPolygons(list(Polygons(list(p), "ID"))) #if (require(rgeos)) { # stplot(sttdf[pol]) # names(sttdf[pol]@traj) # stplot(sttdf[1:2],col=1:2) # stplot(sttdf[,t0]) # stplot(sttdf[,"2013"]) # stplot(sttdf[pol,"2013"]) # is.null(sttdf[pol,t0]) #}
retrieve, or set, information whether time reflects instance (FALSE) or intervals (TRUE)
timeIsInterval(x, ...) timeIsInterval(x) <- valuetimeIsInterval(x, ...) timeIsInterval(x) <- value
x |
object, of any class |
... |
ignored |
value |
logical; sets the timeIsInterval value |
logical; this function sets or retrieves the
attribute timeIsInterval of x, UNLESS x
is of class ST, in which case it sets or retrieves
this attribute for the time slot of the object, i.e.
timeIsInterval(x@time) <- value
From spacetime 0.8-0 on, timeIsInterval is dropped in favour of
a more generic time intervals by specifying endTime of each observation
match two (time) sequences, where each can be intervals or instances.
timeMatch(x, y, returnList = FALSE, ...)timeMatch(x, y, returnList = FALSE, ...)
x |
ordered sequence, e.g. of time stamps |
y |
ordered sequence, e.g. of time stamps |
returnList |
boolean; should a list be returned with all matches (TRUE), or a vector with single matches (FALSE)? |
... |
|
When x and y are of class xts or POSIXct,
end.x and end.y need to specify endpoint of intervals.
In case x and y are both not intervals, matching is
done on equality of values, using match.
If x represents intervals, then the first interval is from
x[1] to x[2], with x[1] included but x[2]
not (left-closed, right-open). In case of zero-width intervals
(e.g. x[1]==x[2]), nothing will match and a warning is raised.
Package intervals is used to check overlap of intervals,
using, interval_overlap.
if returnList = FALSE: integer vector of length
length(x) with indexes of y matching to each of
the elements of x, or NA if there is no match. See section
details for definition of match.
if returnList = TRUE: list of length length(x),
with each list element an integer vector with all the indexes
of y matching to that element of x.
Edzer Pebesma
https://www.jstatsoft.org/v51/i07/
over, timeIsInterval, interval_overlap
t0 = as.POSIXct("1999-10-10") x = t0 +c(0.5+c(2,2.1,4),5)*3600 y = t0 + 1:5 * 3600 x y #timeIsInterval(x) = FALSE #timeIsInterval(y) = FALSE timeMatch(x,y, returnList = FALSE) timeMatch(x,y, returnList = TRUE) #timeIsInterval(y) = TRUE timeMatch(x,y, returnList = FALSE, end.y = delta(y)) timeMatch(x,y, returnList = TRUE, end.y = delta(y)) #timeIsInterval(x) = TRUE timeMatch(x,y, returnList = FALSE, end.x = delta(x), end.y = delta(y)) timeMatch(x,y, returnList = TRUE, end.x = delta(x), end.y = delta(y)) #timeIsInterval(y) = FALSE timeMatch(x,y, returnList = FALSE, end.x = delta(x)) timeMatch(x,y, returnList = TRUE, end.x = delta(x)) x = as.POSIXct("2000-01-01") + (0:9) * 3600 y = x + 1 y[1] = y[2] x y TI = function(x, ti) { timeIsInterval(x) = ti x } #timeMatch(TI(y,FALSE),TI(y,FALSE)) #timeMatch(TI(y,TRUE), TI(y,TRUE)) # #timeMatch(TI(x,FALSE),TI(y,FALSE)) #timeMatch(TI(x,FALSE),TI(y,TRUE)) #timeMatch(TI(x,TRUE), TI(y,FALSE)) #timeMatch(TI(x,TRUE), TI(y,TRUE)) # #timeMatch(TI(x,FALSE),TI(y,FALSE), returnList = TRUE) #timeMatch(TI(x,FALSE),TI(y,TRUE), returnList = TRUE) #timeMatch(TI(x,TRUE), TI(y,FALSE), returnList = TRUE) #timeMatch(TI(x,TRUE), TI(y,TRUE), returnList = TRUE)t0 = as.POSIXct("1999-10-10") x = t0 +c(0.5+c(2,2.1,4),5)*3600 y = t0 + 1:5 * 3600 x y #timeIsInterval(x) = FALSE #timeIsInterval(y) = FALSE timeMatch(x,y, returnList = FALSE) timeMatch(x,y, returnList = TRUE) #timeIsInterval(y) = TRUE timeMatch(x,y, returnList = FALSE, end.y = delta(y)) timeMatch(x,y, returnList = TRUE, end.y = delta(y)) #timeIsInterval(x) = TRUE timeMatch(x,y, returnList = FALSE, end.x = delta(x), end.y = delta(y)) timeMatch(x,y, returnList = TRUE, end.x = delta(x), end.y = delta(y)) #timeIsInterval(y) = FALSE timeMatch(x,y, returnList = FALSE, end.x = delta(x)) timeMatch(x,y, returnList = TRUE, end.x = delta(x)) x = as.POSIXct("2000-01-01") + (0:9) * 3600 y = x + 1 y[1] = y[2] x y TI = function(x, ti) { timeIsInterval(x) = ti x } #timeMatch(TI(y,FALSE),TI(y,FALSE)) #timeMatch(TI(y,TRUE), TI(y,TRUE)) # #timeMatch(TI(x,FALSE),TI(y,FALSE)) #timeMatch(TI(x,FALSE),TI(y,TRUE)) #timeMatch(TI(x,TRUE), TI(y,FALSE)) #timeMatch(TI(x,TRUE), TI(y,TRUE)) # #timeMatch(TI(x,FALSE),TI(y,FALSE), returnList = TRUE) #timeMatch(TI(x,FALSE),TI(y,TRUE), returnList = TRUE) #timeMatch(TI(x,TRUE), TI(y,FALSE), returnList = TRUE) #timeMatch(TI(x,TRUE), TI(y,TRUE), returnList = TRUE)
create table forms of STFDF objects
## S3 method for class 'STFDF' unstack(x, form, which = 1, ...) ## S3 method for class 'STFDF' as.data.frame(x, row.names, ...)## S3 method for class 'STFDF' unstack(x, form, which = 1, ...) ## S3 method for class 'STFDF' as.data.frame(x, row.names, ...)
x |
object of class |
form |
formula; can be omitted |
which |
column name or number to have unstacked |
row.names |
row.names for the data.frame returned |
... |
arguments passed on to the functions unstack or as.data.frame |
unstack returns the data in wide format, with each
row representing a spatial entity and each column a time; see
unstack for details and default behaviour.
as.data.frame returns the data.frame in long format,
where the coordinates of the spatial locations (or line starting
coordinates, or polygon center points) and time stamps are recycled
accordingly.
sp = cbind(x = c(0,0,1), y = c(0,1,1)) row.names(sp) = paste("point", 1:nrow(sp), sep="") library(sp) sp = SpatialPoints(sp) library(xts) time = xts(1:4, as.POSIXct("2010-08-05")+3600*(10:13)) m = c(10,20,30) # means for each of the 3 point locations mydata = rnorm(length(sp)*length(time),mean=rep(m, 4)) IDs = paste("ID",1:length(mydata)) mydata = data.frame(values = signif(mydata,3), ID=IDs) stfdf = STFDF(sp, time, mydata) as.data.frame(stfdf, row.names = IDs) unstack(stfdf) t(unstack(stfdf)) unstack(stfdf, which = 2)sp = cbind(x = c(0,0,1), y = c(0,1,1)) row.names(sp) = paste("point", 1:nrow(sp), sep="") library(sp) sp = SpatialPoints(sp) library(xts) time = xts(1:4, as.POSIXct("2010-08-05")+3600*(10:13)) m = c(10,20,30) # means for each of the 3 point locations mydata = rnorm(length(sp)*length(time),mean=rep(m, 4)) IDs = paste("ID",1:length(mydata)) mydata = data.frame(values = signif(mydata,3), ID=IDs) stfdf = STFDF(sp, time, mydata) as.data.frame(stfdf, row.names = IDs) unstack(stfdf) t(unstack(stfdf)) unstack(stfdf, which = 2)