continuous time spent

[mdsumner]

cut.trip does most of the work, but it seems to add in a stray day with the walrus data.

library(trip)
data(walrus818)
prj <- "+proj=laea +lon_0=-165 +lat_0=90 +ellps=WGS84"
walrus <- spTransform(walrus818, prj)

grid <- rasterize(walrus)

#grid <- rasterize(tr[[1]], method = "density")
tr <- cut(walrus, "1 day")
library(raster)
library(tibble)
library(dplyr)
tib1 <- function(x) {
  bind_cols(as_tibble(as.data.frame(x)), tibble(cell = seq(ncell(x)))) %>% filter(z > 0)
}
st <- bind_rows(lapply(tr, function(x) tib1(rasterize(x, grid = grid))), .id = "interval")
sum(st$z)
[1] 45500400
> cellStats(gg, sum)
[1] 44640000

Somewhere here is integration with splat/SGAT and tabularaster.

[mdsumner]

this works, but a bit slow still

library(trip)
library(dplyr)
library(furrr)
x <- trip(readr::read_csv("inst/extdata/MI_albatross_sub10.csv") %>% dplyr::select(lon, lat, gmt, tag_ID))
x <- spTransform(x, "+proj=laea +lat_0=-90")

r <- raster(x, res = 50000)
spx <- explode(x)
plan(multiprocess)
l <- furrr::future_map_dfr(split(spx, 1:nrow(spx)), function(a) tibble(cell = unlist(cellFromLine(r, a) )), .id = "segment")

l$dur <- spx$timedur[as.integer(l$segment)]
dd <- l %>% group_by(segment) %>% mutate(dur = dur/n()) %>% ungroup() %>% group_by(cell) %>%  summarize(dur = sum(dur))
r[] <- 0
r[dd$cell] <- dd$dur
plot(r)
cellStats(r, sum)
[1] 1286139971
> sum(spx$timedur)
[1] 1286139971