-
Notifications
You must be signed in to change notification settings - Fork 2
/
AT_sdcSpatial_Examples.R
295 lines (230 loc) · 10.2 KB
/
AT_sdcSpatial_Examples.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
#######################################################################
# Script to apply sdcSpatial on 500/1000/2000 m grid count tables
#
# Script applies the following steps (in that order)
#
# - load needed libraries
# - load dummy population data
# - define "raster"-object for grid cells to be used with sdcSpatial
# - define sdc_raster object using population data and raster-object from before
# - check sensitive cells
# - apply protection method; check sensitive cells again and suppress if necessary
# - define focus areas and calculate information loss
#
## -----------------------------------------------------------------------------------------
# load libraries
library(data.table) # r package for data wrangling with large(r) data sets
library(raster) # r package for defining raster-objects -> used by sdcSpatial
library(terra) # used to read in shape file
library(sdcSpatial) # r package to apply sdc methods on spatial data
library(viridis) # r package for color shema (used in plots)
source("helpers.R")
## -----------------------------------------------------------------------------------------
# load population data
load("AT_test_data.RData")
pop_data
# data.table should be one row per person and contain variables
# coordinates for each person (x, y)
# IDs for gridcells
# - L000500 (500m)
# - L001000 (1000m)
# - L002000 (2000m)
# PID (optional)
raster_cells
# data.table containing all raster cells for a country even unpopulated ones
# should contain all grid cell IDs
# - L000500 (500m)
# - L001000 (1000m)
# - L002000 (2000m)
# define cell size
gridsize <- 500 # dimension of rasters in meters
geo <- paste(c("L",paste(rep(0,6-nchar(gridsize))),gridsize),collapse="")
# make table original cell counts
grid_data <- pop_data[,.(count_original=.N),by=c(geo)]
grid_data <- merge(grid_data, raster_cells, by=c(geo),all=TRUE)
# get center coordinates of grid cells - needed later
grid_data[,y:=as.numeric(paste0(substr(get(geo),6,10),"00"))]
grid_data[,x:=as.numeric(paste0(substr(get(geo),12,16),"00"))]
grid_data[is.na(count_original),count_original:=0]
## -----------------------------------------------------------------------------------------
# define "raster"-object for grid cells to be used with sdcSpatial
# download and read in shapefiles
path <- paste0("https://data.statistik.gv.at/data/OGDEXT_RASTER_1_STATISTIK_AUSTRIA_",
geo,"_LAEA.zip")
layer <- paste0("STATISTIK_AUSTRIA_",
geo,"_LAEA")
zipname <- basename(path)
# download .zip
my_shapefile <- paste0(getwd(),"/",zipname)
download.file(path,my_shapefile)
# unzip file
unzipfolder <- paste0(sub(".zip","",my_shapefile))
unzip(my_shapefile,exdir=unzipfolder)
# read shape file
area_shape <- terra::vect(file.path(unzipfolder, paste0(layer,".shp")))
# delete zip file
file.remove(my_shapefile)
# delete folder optional
# unlink(unzipfolder, recursive = TRUE)
# create raster object to be used by sdcSpatial
bounds <- ext(area_shape)
xmax <- bounds[2] + 3*gridsize # 3*gridsize -> include extra boundary on the map
xmin <- bounds[1] - 3*gridsize
ymax <- bounds[4] + 3*gridsize
ymin <- bounds[3] - 3*gridsize
ncol <- (xmax - xmin) / gridsize + 1
nrow <- (ymax - ymin) / gridsize + 1
projection <- sf::st_crs(area_shape)$proj4string
r <- raster::raster(
nrows = nrow,
ncols = ncol,
xmn = xmin - gridsize / 2,
xmx = xmax + gridsize / 2,
ymn = ymin - gridsize / 2,
ymx = ymax + gridsize / 2,
crs = sp::CRS(projection)
)
print(r)
# for overview on projections see:
# https://www.nceas.ucsb.edu/sites/default/files/2020-04/OverviewCoordinateReferenceSystems.pdf
## -----------------------------------------------------------------------------------------
# define sdc_raster object using population data and raster-object from before
# convert populatin data to SpatialPointsDataFrame
pop_data[,value:=1] # set value needed for sdcSpatial
sp_data <- sp::SpatialPointsDataFrame(coords = as.data.frame(pop_data[,.(x,y)]),
proj4string = sp::CRS(projection),
data = as.data.frame(pop_data))
# define sdc_raster object
pop_raster <- sdc_raster(x = sp_data, # SpatialPointsDataFrame holding the data
variable = "value", # variable cannot be missing
r = r, # set specific raster
min_count = 5 # ~ cells should have at least min_count persons
)
# object summary
pop_raster
# mean sensitivity score: share of sensitive cells
pal <- rev(viridis(10))
# plot object
plot(pop_raster,
value="count", # what value should be plotted
col = pal)
##
# check sensitive cells
sensitive_cells <- is_sensitive(pop_raster) # get sensitive cells
data_sens <- as.data.frame(sensitive_cells,xy=TRUE) # convert to data frame
setDT(data_sens)
# compare with input data (sanity check)
grid_data[,sensitive_direct:=count_original<5 & count_original>0]
grid_data_check <- merge(grid_data,data_sens,by=c("x","y"),all=TRUE)
# filter by all grid cells which are part of country
# use shape file for this
grid_data_check[,part_of_country:=get(geo) %in% area_shape$ID]
grid_data_check[,.N,by=.(sensitive,sensitive_direct,part_of_country)]
## -----------------------------------------------------------------------------------------
# apply protection method; check sensitive cells again and suppress if necessary
# lets compare the following 3 different methods
# protect_quadtree()
# protect_remove()
# protect_smooth()
# apply removal of sensitive cells
protected_rm <- remove_sensitive(x = pop_raster, min_count = 5)
print(protected_rm)
# apply quadtree
protected_qt <- protect_quadtree(x = pop_raster, min_count = 5, max_zoom = 2)
print(protected_qt)
# apply smoothing
protected_smooth_500 <- protect_smooth(x = pop_raster, bw = 500)
print(protected_smooth_500)
## -----------------------------------------------------------------------------------------
# define focus areas and calculate info loss for them
# set coordinates of center and radius for each focus area
focus_areas <- list(
Area1 = list(center = c(x = 4790500, y = 2809500),
radius = 26000),
Area2 = list(center = c(x = 4547500, y = 2747000),
radius = 15000),
Area3 = list(center = c(x = 4298000, y = 2707500),
radius = 14500)
)
protected_grids <- list(remove=protected_rm,
quadtree = protected_qt,
smooth500 = protected_smooth_500)
# focus areas for unprotected maps
sdcObj_orig <- lapply(focus_areas,function(z,pop_ras){
focus_sdcSpatialObj(pop_ras,focus=z)
},pop_ras=pop_raster)
# focus areas for protected maps
sdcObj_focus1 <- lapply(protected_grids,focus_sdcSpatialObj, focus=focus_areas$Area1)
sdcObj_focus2 <- lapply(protected_grids,focus_sdcSpatialObj, focus=focus_areas$Area2)
sdcObj_focus3 <- lapply(protected_grids,focus_sdcSpatialObj, focus=focus_areas$Area3)
sdcObj_focus1
sdcObj_focus2
sdcObj_focus3
plot(sdcObj_orig[[1]], value = "count", col = pal)
plot(sdcObj_focus1$remove, value = "count", col = pal)
plot(sdcObj_focus1$quadtree, value = "count", col = pal)
plot(sdcObj_focus1$smooth, value = "count", col = pal)
plot(sdcObj_orig[[2]], value = "count", col = pal)
plot(sdcObj_focus2$remove, value = "count", col = pal)
plot(sdcObj_focus2$quadtree, value = "count", col = pal)
plot(sdcObj_focus2$smooth, value = "count", col = pal)
plot(sdcObj_orig[[3]], value = "count", col = pal)
plot(sdcObj_focus3$remove, value = "count", col = pal)
plot(sdcObj_focus3$quadtree, value = "count", col = pal)
plot(sdcObj_focus3$smooth, value = "count", col = pal)
## -----------------------------------------------------------------------------------------
# Calculate HD and KWD for focus areas
sdcObj_focus <- list(Area1 = sdcObj_focus1,
Area2 = sdcObj_focus2,
Area3 = sdcObj_focus3)
# get raster from original input
original_raster <- pop_raster$value[["count"]]
v_o <- raster::getValues(original_raster)
v_o[is.na(v_o)] <- 0
xy <- raster::xyFromCell(original_raster, 1:raster::ncell(original_raster))
# transform coordinates to integer
xy_int <- copy(xy)
xy_int[,"x"] <- as.numeric(factor(xy_int[,"x"]))
xy_int[,"y"] <- as.numeric(factor(xy_int[,"y"]))
out <- list()
# iterate over focus areas
for(f in names(sdcObj_focus)){
x_center <- focus_areas[[f]]$center[["x"]]
y_center <- focus_areas[[f]]$center[["y"]]
r_focus <- focus_areas[[f]]$radius
# get position of center and radius for integer transformed coordinates
int_center <- xy_int[xy[,"x"]==x_center & xy[,"y"]==y_center]
int_radius <- r_focus/gridsize
# get filter for focus area
xy_filter <- xy[,"x"] %between% c(x_center-r_focus,x_center+r_focus) &
xy[,"y"] %between% c(y_center-r_focus,y_center+r_focus)
# filter for focus area + 5*gridsize borders
r_focus_ext <- r_focus+5*gridsize
xy_filter_kwd <- xy[,"x"] %between% c(x_center-r_focus_ext,x_center+r_focus_ext) &
xy[,"y"] %between% c(y_center-r_focus_ext,y_center+r_focus_ext)
# iterarte over protection methods
for(m in names(sdcObj_focus[[f]])){
protected_raster <- protected_grids[[m]]$value[["count"]]
v_x <- raster::getValues(protected_raster)
v_x[is.na(v_x)] <- 0
# v_x[xy_filter==FALSE] <- 0 # needed for KWDSpatial::focusArea()
sum(v_o[xy_filter_kwd==TRUE & xy_filter==TRUE])
sum(v_x[xy_filter_kwd==TRUE & xy_filter==TRUE])
# calculate HD
v_o_hd <- v_o[xy_filter==TRUE]
v_x_hd <- v_x[xy_filter==TRUE]
hd <- (1/sqrt(2)) * sqrt((sum(sqrt(v_o_hd/sum(v_o_hd)) - sqrt(v_x_hd/sum(v_x_hd)))^2))
# calculate KWD
kwd <- SpatialKWD::focusArea(Coordinates = xy_int[xy_filter_kwd,],
Weights = cbind(v_o[xy_filter_kwd],v_x[xy_filter_kwd]),
x = int_center[1],
y = int_center[2],
radius = int_radius,
method = "exact", area = "Inf")
out_i <- data.table(Area=f,Method=m,hd=hd,kwd=kwd$distance/sum(v_x_hd))
out <- c(out,list(out_i))
}
}
out <- rbindlist(out)
out
# for 0 distance for KWD see current issue also raised in Example_spatialKWD_focusAreas.Rmd