stconstruct: create ST* objects from long or wide tables

In spacetime: Classes and Methods for Spatio-Temporal Data

create ST* objects from long or wide tables

Description

create ST* objects from long or wide tables

Usage

 
stConstruct(x, space, time, SpatialObj = NULL, TimeObj = NULL, 
	crs = CRS(as.character(NA)), interval, endTime)

Arguments

x	object of class matrix or data.frame, holding the long, space-wide or time-wide table; see details.
space	in case x is a long table, character or integer holding the column index in x where the spatial coordinates are (if length(space)==2) or where the ID of the spatial location is (if (length(space)==1). If x is a space-wide table, a list with each (named) list element a set of columns that together form a variable
time	in case x is a long table, character or integer indicating the column in x with times;
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 space
crs	object of class CRS-class; only used when coordinates are in x and no CRS can be taken from SpatialObj
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 POSIXct, specifying (if present) the end points of observation time intervals

Details

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.

Value

Depending on the arguments, an object of class STIDF or STFDF.

References

https://www.jstatsoft.org/v51/i07/

Examples

# 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)
}

spacetime documentation built on Sept. 12, 2024, 7:42 a.m.