Anonymize spatial data

Question

I'm putting together some training materials for a workshop on R for spatial data analysis in archaeology, and need to anonymize the true site locations (confidential information) while keeping the spatial integrity of the features within the site's bounding box.

My initial thought is to subtract the minimum X & Y values of the bounding box (all are in UTM coords) from all of the geometries - i.e., make the whole thing an arbitrary relative grid.

I have ESRI point and polygon shapefiles imported to R with sf, and some rasters loaded with terra but not sure how to go about doing the global spatial adjustment for either.

I also use QGIS, so some solution through there to pre-process the geometries before loading to R would work as well - but again, don't know how to do that.

Answer 1

Here is how you can do that.

Original data:

library(terra)
r <- rast(system.file("ex/elev.tif", package="terra"))
v <- vect(system.file("ex/lux.shp", package="terra"))
p <- spatSample(v, 10)

plot(r)
lines(v)
points(p)

Shift and rotate

# move to different location
rs <- shift(r, -10, 10)
vs <- shift(v, -10, 10)
ps <- shift(p, -10, 10)

# rotate
rst <- t(rs)
e <- ext(rst)
x0 = e[2] + (e[2]-e[2])/2
x1 = e[3] + (e[4]-e[3])/2
vst <- t(vs) |> spin(180, x0, x1)
pst <- t(ps) |> spin(180, x0, x1)

plot(rst)
lines(vst)
points(pst)

Of course if you know the true location of some places on the map, e.g. if you know this is Luxembourg, it is not too difficult to reverse engineer and get the original data.

You could also add some jitter to the points:

xy <- crds(pst)
xy <- xy + runif(length(xy), -0.05, 0.05)
pstj <- vect(xy)

Set a fake coordinate reference system to all the datasets, like

crs(pstj) <- "+proj=utm +zone=23"

And save the data like this

writeVector(pstj, "sites.shp")
writeRaster(rst, "grid.shp")