Short moveVis example on how to overlay frames with additional transparent rasters changing over time

moveVis_overlay_base_map_with_transparent_rasters.R

library(moveVis)
library(move)
library(ggplot2)

data("move_data", "basemap_data")
m <- align_move(move_data, res = 4, unit = "mins")

## create frames using mapbox base map
frames <- frames_spatial(m, map_type = "satellite", map_service = "mapbox", map_token = "YOUR_VERY_LONG_TOKEN",
                         path_legend = T)

r_list <- basemap_data[[1]]
r_times <- basemap_data[[2]]

# crop to plot extent
r_list <- lapply(r_list, raster::crop, y = extent(m), snap = "out")

# here I fake a 3 class categorical raster
r_list <- lapply(r_list, function(x){
  y <- setValues(x, round(getValues(x)*3))
  y[y == 0] <- 3
  return(y)
})

## first, we need a list of rasters with the same length as frames
## and with each raster at the correct time in that list. For this:

# create timestamps of m
ts <- sort(unique(timestamps(m))) 
# ts has same length as frames, since frames were created from the timestamps of m
# caclulate difftime for each time stamp to all time stamps of your rasters and get minimum absolute diff
r_pos <- sapply(ts, function(x) which.min(abs(difftime(x, r_times))))
# the subscript with min abs time difference between the raster time and the times of m
# is what we used to assign the rasters per time:
r_list <- lapply(r_pos, function(i) r_list[[i]])
# now, rasters are assigned per frame time


## just call add_gg with a ggplot2 raster drawing command,

# for example the ggR function in RStoolbox in the layer mode with 30 % opacity:
frames_overlayed <- add_gg(frames, gg = expr(RStoolbox::ggR(data, alpha = 0.8, forceCat = T, ggLayer = T)), data = r_list)
# iterates over all frames and over r_list (same lengths) and applies the ggR function
# instead of ggR, you could use any ggplot2 expression you like to create a raster layer
frames_overlayed[[100]]

# or if you want so specify colours and alpha per class ( e.g. make a class be completely transparent),
# you can use geom_raster.
# turn r_list into data.frames for ggplot2 and define colours and alphas per value (here we have 3 classes)
r_df <- lapply(r_list, function(x) data.frame(coordinates(x),
                                              vals = values(x),
                                              cols = plyr::mapvalues(values(x), 1:3, c("darkgreen", "green", "blue")),
                                              alpha = plyr::mapvalues(values(x), 1:3, c(0, 0.4, 1))))
# and use geom_raster with r_df to get colours and alpha per class
frames_overlayed <- add_gg(frames, gg = expr(geom_raster(data = r_df[[100]], mapping = aes(x = x, y = y, fill = cols, alpha = alpha), show.legend = F)), data = r_df)
frames_overlayed[[100]]

# additional note:
# With geom_raster, plot extents are changed, while with ggR using annotation_raster, plot extents remain
# so either crop to the correct extent or fixate the ggplot axis limits of frames before adding geom_raster
# I did not manage to display a legend for fill of geom_raster and do not know why, but had no time to investigate.
# alpha legend is displayed though, if show.legend = T