-
Notifications
You must be signed in to change notification settings - Fork 0
/
nycGridLib.R
407 lines (321 loc) · 12.4 KB
/
nycGridLib.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
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
#
# Library for developing derivatives and plots from gridded data
#
# Grid extents
# CellNumbers(GridExtents, N)
# Testing:
# setwd("C:/Users/mhdan_000/Dropbox/Incubator/nyctaxi/")
# From census map extent (nycMapping.R)
# min.x = min(nyc_map$lon)
# max.x = max(nyc_map$lon)
# min.y = min(nyc_map$lat)
# max.y = max(nyc_map$lat)
# Testing
# setwd("C:/Users/mhdan_000/Dropbox/Incubator/nyctaxi/")
censusExtent <- c(-74.25559,-73.70001, 40.49612, 40.9153) # c(min.x, max.x, min.y, max.y)
latLonToGrid <- function(lat, lon, n = 100, offset = 0.00001, leftToRight = T){
# Returns the cell number corresponding to a lat/lon
# in a roughing grid (n x n), within the extents of the nyc_map.
# Vectorized
# Construct extremes of nyc_map (for use in roughing bounds)
lat.min = censusExtent[3] - offset # eliminates edge case (1,0] in roughingGrid()
lat.max = censusExtent[4]
lon.min = censusExtent[1] - offset # eliminates edge case (1,0] in roughingGrid()
lon.max = censusExtent[2];'.;;;;;;;;;;;;;;;'
lat.width = (lat.max - lat.min) / n
lon.width = (lon.max - lon.min) / n
# For a point, returning the grid cell number
# (top to bottom, *LEFT TO RIGHT*) ro
# NOTE: Prior versions were RIGHT TO LEFT, such as nycMapping.roughingGrid(). this has been amended
if(leftToRight){
grid_id = (lat.max - lat) %/% lat.width * n +
(lon - lon.min) %/% lon.width + 1
} else {
grid_id = (lat.max - lat) %/% lat.width * n +
(lon.max - lon) %/% lon.width + 1
}
# Clipping index
na_index = lon > lon.max |
lat > lat.max |
lon < lon.min |
lat < lat.min
grid_id[na_index] = NA # Not in place... could be problematic for larger vectors
return(grid_id)
# Testing
# Check corners
# latLonToGrid(lat = nyc_map$lat, lon = nyc_map$lon, n = 20)
# Time it: 0.6s with boundary checking, 0.35s w/o
# system.time(dt200901[,grid := roughingGrid(pickup_longitude, pickup_latitude)])
# sum(is.na(dt200901$grid)) # 247k outside the grid
}
gridHFlip <- function(grid_id, n){
# Fixes prior grid assignments by flipping horizontally the raster grid_ids
# Mispecifying n will mess up the assignment.
# Get colNum (RIGHT to LEFT, but shall be swapped, so irrelevant)
colNum = (grid_id - 1) %% n + 1
# Reflect grid about center by adding the reflector vector indexed on col#
reflector = seq((n-1), (1-n), by = -2)
return(grid_id + reflector[colNum])
# Testing
# gridHFlip(18:1, n = 9)
# gridHFlip(1:20, n = 9)
# matrix(gridHFlip(1:9, 3), ncol = 3,byrow = T)
# matrix(gridHFlip(1:16, 4), ncol = 4,byrow = T)
}
gridReduce <- function(grid_id, n = 100, smaller_n = 25){
# Grid downsize to finer mesh
# Takes a grid_id and linear grid number
# Outputs a grid_id for a new new, coarser mesh defined by "smaller_n" the linear size
# Vectorized
#
# Idea:
# Get col/row of fine mesh
# downsize to col/row of new mesh by fine/coarse factor
# Assign new id
#
reduceFactor = n / smaller_n
if(reduceFactor != floor(reduceFactor)){
print(paste(n, "is not an integer multiple of", smaller_n))
return(NULL)
}
# Zero indexed
fineCol = (grid_id - 1) %% n
fineRow = grid_id %/% n
coarseCol = fineCol %/% reduceFactor
coarseRow = fineRow %/% reduceFactor
# One indexed
coarse_grid_id = 1 + coarseCol + coarseRow * smaller_n
return(coarse_grid_id)
}
getCentroidGrid <- function(grid_id, n, offset = 0.00001, leftToRight=T){
# Given grid_ids and linear grid size
# return centroids of grid points
# Zero Indexed row/col:
row = (grid_id - 1) %/% n
col = (grid_id - 1) %% n
lat.width = (offset + censusExtent[4] - censusExtent[3]) / n
lon.width = (offset + censusExtent[2] - censusExtent[1]) / n
# centroid of Upper-left origin
# for standard top-bottom, left-to-right raster cell ordering
# x, y
# lon, lat
origin = c(censusExtent[1], censusExtent[4]) + 1/2 * c(lon.width, -lat.width)
centroid.lat = origin[2] - row * lat.width
centroid.lon = origin[1] + col * lon.width
df = data.frame(lon = centroid.lon, lat = centroid.lat, grid_id = grid_id)
coordinates(df) = ~ lon + lat
proj4string(df) = CRS("+proj=longlat +datum=WGS84")
return(df)
# Test: getCentroidGrid(1, 100)
}
gridToGraphOld <- function(grid_id,n){
library(plyr)
connected <- function(grid_id_1, grid_id_2, n){
if(abs((grid_id_1 - 1) %% n - (grid_id_2 - 1) %% n) == 1 &
(grid_id_1 - 1) %/% n == (grid_id_2 - 1) %/% n){
return(1)
}
if(abs((grid_id_1 - 1) %/% n - (grid_id_2 - 1) %/% n) == 1 &
(grid_id_1 - 1) %% n == (grid_id_2 - 1) %% n){
return(1)
}
return(0)
}
# Cartesian product: grid_id X grid_id
df = expand.grid(gid_1 = 1:n^2, gid_2 = 1:n^2)
# For each combo, check connected(), symmetry => matrix fill direction irrelevant
M = matrix(mapply(connected, df$gid_1, df$gid_2, n), ncol = n^2, nrow = n^2)
# For grids not in the provided grid_ids, delete adjacency
del_grids = setdiff(1:n^2, unique(grid_id))
M[del_grids, ] = 0
M[, del_grids] = 0
}
gridToGraph <- function(grid_id, n){
# Given a vector of grid_ids present in an (n x n) grid,
# return the adjacency matrix for the undirected graph
# Idea 1: generate adjacency for square matrix,
# wipe out missing (non-physical) cells
#
subdiag <- function(vec, size, offset=0){
# diag(x,k) matlab equivalent
# https://stackoverflow.com/questions/7745363/r-equivalent-to-diagx-k-in-matlab
M <- matrix(0, size, size)
M[row(M)-offset == col(M)] <- vec
return(M)
}
grid_length = n^2
M = subdiag(1, grid_length, 1) +
subdiag(1, grid_length, -1) +
subdiag(1, grid_length, n) +
subdiag(1, grid_length, -n)
M[row(M) %% n == 0 & col(M) %% n == 1] = 0
M[row(M) %% n == 1 & col(M) %% n == 0] = 0
# For grids not in the provided grid_ids, delete adjacency
del_grids = setdiff(1:n^2, unique(grid_id))
M[del_grids, ] = 0
M[, del_grids] = 0
return(M)
# Test: gridToGraph(grid_id = 1:9, n = 3)
}
# M = gridToGraph(grid_id = 1:10000, n = 100)
graphWater <- function(graph){
# Using the extent, the square graph, and a map,
# null out the graph connections to water/uninhabitable areas
# Idea:
# For each point associated with the graph, check if it's in NY or NJ.
# If not, zero out the graph connections for that row/column
library(rgdal)
library(maptools)
nytracts = spTransform(readOGR("./nyct2010_15b", layer = "nyct2010"), CRS("+proj=longlat +datum=WGS84"))
njtracts = spTransform(readOGR("./Census2010Tr2012", layer = "Govt_TIGER2012_tract2010"), CRS("+proj=longlat +datum=WGS84"))
boroughs = unionSpatialPolygons(nytracts, nytracts$BoroCode)
nj = unionSpatialPolygons(njtracts, njtracts$COUNTYFP)
n = sqrt(dim(graph)[1])
# Test
# centroids = getCentroidGrid(1:100, 10)
centroids = getCentroidGrid(1:n^2, n)
remove = is.na(over(centroids, nj)) & is.na(over(centroids, boroughs))
graph[remove,] = 0
graph[,remove] = 0
return(graph)
}
graphConnect <- function(graph, gids1, gids2){
# Connect the given graph ids on the graph
# operation is copy, not in-place
# mapply(function(gid1,gid2){
# graph[gid1, gid2] = 1
# graph[gid2, gid1] = 1
#}, gids1, gids2)
for(i in 1:length(gids1)){
graph[gids1[i], gids2[i]] = 1
graph[gids2[i], gids1[i]] = 1
}
return(graph)
}
getBridgesAndTunnels <- function(){
# Read bridges and tunnels file
# Return df.bridges = data.frame(w_lat, w_lon, e_lat, e_lon, name) row pairs in DF
# Must be in project src dir
return(read.csv('bridgesAndTunnels.csv'))
}
graphBridgesAndTunnels <- function(graph){
# Take bridgesAndTunnels,
# append graph according to entrances/exists of bridges and tunnels
bAndT = getBridgesAndTunnels()
n = sqrt(dim(graph)[1])
bAndT$w_grid_id = latLonToGrid(lat = bAndT$w_lat, lon = bAndT$w_lon, n = n)
bAndT$e_grid_id = latLonToGrid(lat = bAndT$e_lat, lon = bAndT$e_lon, n = n)
bAndT$connect = bAndT$w_grid_id != bAndT$e_grid_id
print(bAndT)
return(graphConnect(graph, gids1 = bAndT$e_grid_id[bAndT$connect], gids2= bAndT$w_grid_id[bAndT$connect]))
}
plotGraph <- function(graph){
# Given a graph, plot the connections as lines
library(plyr)
n = sqrt(dim(graph)[1]) # assumes graph represents square grid
# Get list of connected graph ids, plot the connected nodes in the upper triangular (symmetry)
line_grid_ids = which(graph * upper.tri(graph) == 1, arr.ind = T)
# Make data frame of pairs of points
c1 = getCentroidGrid(line_grid_ids[,1], n)
c2 = getCentroidGrid(line_grid_ids[,2], n)
df = as.data.frame(cbind(coordinates(c1), coordinates(c2)))
names(df) <- c("p1_lon", "p1_lat", "p2_lon", "p2_lat")
# plot lines
plot(1, main="Transportation Graph",type="n", xlab="", ylab="", xlim = censusExtent[1:2], ylim = censusExtent[3:4])
m_ply(.data = df, .fun = function(p1_lon, p1_lat, p2_lon, p2_lat){
lines(c(p1_lon, p2_lon), c(p1_lat, p2_lat))
})
}
# testGraph = gridToGraph(1:25, 5)
# testGraph2 = graphWater(testGraph)
# N = gridToGraph(grid_id = 1:2500, n = 50)
# N.wat = graphWater(N)
# N.bat = graphBridgesAndTunnels(N.wat)
# plotGraph(N.bat)
M = gridToGraph(grid_id = 1:10000, n = 100)
M.wat = graphWater(M)
A = graphBridgesAndTunnels(M.wat) # Adjacency matrix A
saveAdjacencyMatrix <- function(A, filename = "./"){
}
saveGraphPlot <- function(graph){
png("./images/transport_graph.png", width = 800, height = 800)
plotGraph(A)
dev.off()
}
# saveGraphPlot(A)
connectivity = colSums(A) # Number of connected grid points
diffuse <- function(V, delta=0){
averages = colSums(V * A)/connectivity
averages[is.na(averages)] = 0
del.V = (averages-V) + delta
return(V + del.V)
}
getGridPolys <- function(n = 100, offset = 0.00001){
# Returns spatialpolygonsdataframe corresponding to the grid
# Grid ID order is raster standard: Top to bottom outter, Left to Right inner
library(rgeos)
library(data.table)
lat.min = censusExtent[3] - offset # eliminates edge case (1,0] in roughingGrid()
lat.max = censusExtent[4]
lon.min = censusExtent[1] - offset # eliminates edge case (1,0] in roughingGrid()
lon.max = censusExtent[2]
lat.width = (lat.max - lat.min) / n
lon.width = (lon.max - lon.min) / n
n.grids = n*n
grids = 1:n.grids
# generate origin of upper-left corners
# Zero indexed:
row = (grids - 1) %/% n
column = (grids - row * n) - 1
origins = data.table(lon = lon.min + column * lon.width, lat = lat.max - row * lat.width, id = grids)
# Generate polygon for every grid
Mxy <- function(x,y){
o = c(x,y)
matrix(c(o, o + c(0, -lat.width), o + c(lon.width, -lat.width), o + c(lon.width, 0), o),ncol = 2, byrow = T)
}
sp = origins[, list(rects = list(Mxy(lon, lat))), by = list(id)]
# Create SP
polys <- SpatialPolygons(mapply(function(rect, id) {
Polygons(list(Polygon(rect, hole = F)), ID=id)
}, as.list(sp$rects), as.list(sp$id)))
# Verify plotting from upper left
# plot(getGridPolys[c(1:10, 86:100)])
proj4string(polys) <- CRS("+proj=longlat +datum=WGS84")
return(polys )
}
ps = getGridPolys()
library(ggplot2)
fps = fortify(ps)
plotGridData <- function(fortified_polys, V, title="", range.clip=10){
# Plots data in the grid, must be ordered by grid ID,
# must be nxn or length(V) == n^2
library(ggplot2)
p <-ggplot() + geom_polygon(data = fortified_polys, alpha = 0.9, color = NA,
aes(x = long, y = lat, group = id, fill = V[as.integer(id)]),
size = 0) + ggtitle(title) +
scale_fill_gradient2(limits = c(-range.clip,range.clip), high = "blue", mid = "white", low = "red") +
coord_map(xlim = censusExtent[1:2], ylim = censusExtent[3:4]) +
theme(legend.position = "bottom",
axis.line=element_blank(),
axis.text.x=element_blank(),
axis.text.y=element_blank(),
axis.ticks=element_blank(),
axis.title.x=element_blank(),
axis.title.y=element_blank()) +
labs(fill = "Density")
plot(p)
}
###############################################################################
# Diffusion model, proof of concept
###############################################################################
V <- as.vector(matrix(c(rep(10, 5000), rep(-10,5000)), ncol=100, nrow=100, byrow = T))
L <- list(V)
for(i in 2:100){
L[[i]] <- diffuse(L[[i-1]])
}
png("./images/transport_example_002.png", width = 800, height = 800)
plotGridData(fps, L[[2]], title = "Time 002")
dev.off()
png("./images/transport_example_100.png", width = 800, height = 800)
plotGridData(fps, L[[100]], title = "Time 100")
dev.off()