-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathstations-classification.Rmd
562 lines (469 loc) · 23.7 KB
/
stations-classification.Rmd
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
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
---
output: github_document
---
The aim of the code in this document is to classify stations according to the demographics of people living near them.
# Setup
```{r}
library(lchs)
tmap_mode("view")
knitr::opts_chunk$set(eval = TRUE)
```
```{r, eval=FALSE}
install.packages("showtext")
library(showtext)
font_add(family = "Avenir Book", regular = "~/hd/backups/Avenir Book/Avenir Book.ttf")
font_add_google("Gochi Hand", "gochi")
font_add_google("Schoolbell", "bell")
showtext_auto()
x11()
hist(rnorm(1000), breaks = 30, col = "steelblue", border = "white",
main = "", xlab = "", ylab = "")
title("Histogram of Normal Random Numbers", family = "bell", cex.main = 2)
title(ylab = "Frequency", family = "gochi", cex.lab = 2)
text(2, 70, "N = 1000", family = "bell", cex = 2.5)
```
# Input data
Get, plot and preprocess stations data
```{r, eval=TRUE}
# from bikedata stations table
stations_region = readd(stations_region)
stations = readd(stations_yearly)
stations$id = as.character(stations$ucl_id)
stations_2019 = stations %>% filter(year == 2019)
qtm(stations_region) +
qtm(stations_2019)
```
# Change in income decile of LSOA over time
```{r}
# see code that generates this below
lsoa = read_sf("lsoa_bikeshare.geojson")
plot(lsoa)
lsoa = lsoa %>% select(`Income decile` = Income.Decile..where.1.is.most.deprived.10..of.LSOAs.)
plot(lsoa)
stations_yearly_lsoa = st_join(stations, lsoa)
saveRDS(stations_yearly_lsoa, "stations_yearly_lsoa.Rds")
# counts
stations_yearly_lsoa %>%
filter(year %in% c(2010, 2012, 2014, 2019)) %>%
ggplot(aes(`Income decile`)) +
geom_bar(fill = "blue") +
facet_wrap(~year) +
ggthemes::theme_clean(base_family = "gochi")
# props
# source("get-imd-data.R")
source("code/old-code/get-imd-data.R") # if running from root dir
stations_yearly_lsoa %>%
filter(year %in% c(2010, 2012, 2014, 2019)) %>%
# mutate(n_year = sum(year == year)) %>% View()
ggplot(aes(`Income decile`)) +
geom_bar(aes(y=..count../sum(..count..)), fill = "blue") +
scale_y_continuous(labels = scales::percent_format()) +
facet_wrap(~year) +
ggthemes::theme_clean()
stations_yearly_lsoa %>%
filter(year %in% c(2010, 2012, 2014, 2019)) %>%
# mutate(n_year = sum(year == year)) %>% View()
ggplot(aes(`Income decile`)) +
geom_bar(data = lsoa_lnd_joined, aes(`Income decile`, y=..count../sum(..count..)), fill = "grey", width = .95) +
geom_bar(aes(y=..count../sum(..count..)), fill = "blue", width = .60, alpha = 0.7) +
scale_y_continuous(labels = scales::percent_format()) +
facet_wrap(~year)
stations_yearly_4 = stations_yearly_lsoa %>%
filter(year %in% c(2010, 2012, 2014, 2019)) %>%
sf::st_drop_geometry() %>%
group_by(year) %>%
mutate(n_year = n()) %>%
ungroup() %>%
group_by(year, `Income decile`) %>%
summarise(
n_decile_year = n(),
n_year = unique(n_year),
s = diff(range(n_year)),
) %>%
mutate(p = n_decile_year / n_year) %>%
group_by(year) %>%
mutate(t = sum(p))
# stations_yearly_4 = stations_yearly_4 %>%
# mutate(`Income decile` = formatC(x = `Income decile`, width = 2, flag = "0"))
stations_yearly_4 = stations_yearly_4 %>%
mutate(`Income decile` = as.numeric(x = `Income decile`))
stations_yearly_4 %>% ggplot() +
geom_bar(data = lsoa_lnd_joined, aes(`Income decile`, y=..count../sum(..count..) * 4),
fill = "#bdbdbd", width = .95) +
geom_bar(aes(`Income decile`, p), stat = "identity", fill = "#3182bd", width = .60, alpha = 0.7) +
scale_y_continuous(labels = scales::percent_format()) +
scale_x_continuous(breaks = 1:10) +
facet_wrap(~year) +
ylab("Proportion of docking stations in each income decile")
# 1 row version
stations_yearly_4 %>% ggplot() +
geom_bar(data = lsoa_lnd_joined, aes(`Income decile`, y=..count../sum(..count..) * 4),
fill = "#bdbdbd", width = .95) +
geom_bar(aes(`Income decile`, p), stat = "identity", fill = "#3182bd", width = .60, alpha = 0.7) +
scale_y_continuous(labels = scales::percent_format()) +
scale_x_continuous(breaks = 1:10) +
facet_wrap(~year, nrow = 1) +
ylab("% stations in each income decile") +
theme_minimal() +
theme(
axis.text.y = element_text(hjust = 2, margin = margin(l = -30)),
axis.title.y = element_blank()
) +
scale_y_continuous(breaks = seq(0, 0.25, by = 0.05),
labels = c("", "", "10%", "15%", "20%", "25%"))
# annotate("text", label = "25%", x = 1, y = .25)
# theme_set(ggthemes::theme_clean(base_family = "Avenir Book"))
# theme(axis.title=element_blank(), panel.border = element_blank(),
# strip.text.y = element_text(angle=0))
ggsave("figures/stations-imd-facet-4-grey.png", width = 6, height = 3)
magick::image_read("figures/stations-imd-facet-4-grey.png")
# calculate scores
stations_income_yearly_df = stations_yearly_lsoa %>%
group_by(year) %>%
summarise(
median_income_decile = median(`Income decile`),
mean_income_decile = mean(`Income decile`),
sd_income_decile = sd(`Income decile`)
) %>%
st_drop_geometry()
stations_income_yearly_df
# ggsave("figures/stations-imd-facet-4.png")
```
# Join with stations data
```{r}
# trips_df = fst::read_fst("trips_df_all.fst")
trips_df = readd(trips_df)
stations = readd(stations)
stations = st_join(stations, lsoa)
stations$years_in_operation = as.numeric(max(trips_df$start_time) - stations$created_dt + 100) / 365
trips_by_origin_station = trips_df %>%
group_by(id = start_station_id) %>%
summarise(total_n_trips_start = n())
trips_by_origin_station$id = as.character(trips_by_origin_station$id)
head(stations$ucl_id)
head(trips_by_origin_station$id)
stations = inner_join(stations %>% mutate(id = as.character(ucl_id)), trips_by_origin_station)
summary(stations$total_n_trips_start) # all stations there!
hist(stations$total_n_trips_start)
stations$trips_per_year = stations$total_n_trips_start / stations$years_in_operation
tm_shape(stations %>% select(operator_name, trips_per_year)) +
tm_dots(size = "trips_per_year", alpha = 0.5) + # incorrect labels - bug in tmap?
tm_scale_bar()
summary(stations$trips_per_year)
plot(stations$trips_per_year) # 500000 seems excessive
plot(stations$years_in_operation, stations$trips_per_year)
stations$years_in_operation[stations$trips_per_year > 50000 & stations$years_in_operation < 2] # all seem suspect
median(stations$years_in_operation)
# stations$years_in_operation[stations$trips_per_year > 50000 & stations$years_in_operation < 2 ] = median(stations$years_in_operation)
stations$years_in_operation[stations$trips_per_year > 50000 & stations$years_in_operation < 2 ] = NA
stations$trips_per_year = stations$total_n_trips_start / stations$years_in_operation
tm_shape(stations %>% select(operator_name, trips_per_year)) +
tm_dots(size = "trips_per_year", alpha = 0.5) + # incorrect labels - bug in tmap?
tm_scale_bar() # looks good!
tm_shape(stations %>% select(operator_name, total_n_trips_start)) +
tm_dots(size = "total_n_trips_start", alpha = 0.5) + # incorrect labels - bug in tmap?
tm_scale_bar()
# tm_shape(stations) + tm_markers(text = "operator_name")
mapview::mapview(stations)
```
Add additional usage stats
```{r, eval=FALSE}
library(lubridate)
am_peak_int <- interval(hms::as_hms("06:00:00"), hms::as_hms("09:59:59"))
pm_peak_int <- interval(hms::as_hms("16:00:00"), hms::as_hms("19:59:59"))
trips_am_peak = trips_df %>% filter(as.POSIXct(hms::as_hms(start_time)) %within% am_peak_int)
trips_pm_peak = trips_df %>% filter(as.POSIXct(hms::as_hms(start_time)) %within% pm_peak_int)
fst::write_fst(trips_am_peak, "trips_am_peak.fst")
fst::write_fst(trips_pm_peak, "trips_pm_peak.fst")
piggyback::pb_upload("trips_am_peak.fst")
piggyback::pb_upload("trips_pm_peak.fst")
```
## Starting from am/pm peak data
```{r}
trips_pm_peak = fst::read.fst("trips_pm_peak.fst")
trips_am_peak = fst::read.fst("trips_am_peak.fst")
# am
trips_by_origin_station = trips_am_peak %>%
group_by(id = start_station_id) %>%
summarise(total_n_am_peak = n())
trips_by_origin_station$id = as.character(trips_by_origin_station$id)
stations = inner_join(stations, trips_by_origin_station)
stations$trips_per_year_am = stations$total_n_am_peak / stations$years_in_operation
# pm
trips_by_origin_station = trips_pm_peak %>%
group_by(id = start_station_id) %>%
summarise(total_n_pm_peak = n())
trips_by_origin_station$id = as.character(trips_by_origin_station$id)
stations = inner_join(stations, trips_by_origin_station)
stations$trips_per_year_pm = stations$total_n_pm_peak / stations$years_in_operation
sf::write_sf(stations, "stations_clean_am_pm.geojson", delete_dsn = TRUE)
piggyback::pb_upload("stations_clean_am_pm.geojson")
```
```{r}
tmap_mode("plot")
m = tm_shape(stations %>% select(operator_name, trips_per_year, trips_per_year_am, trips_per_year_pm, years_in_operation)) +
tm_dots(size = c("trips_per_year_am", "trips_per_year_pm"),
col = "years_in_operation", palette = "Spectral", contrast = c(0, 0.8), alpha = 0.5, scale = 1.5, title.size = "") +
tm_layout(panel.show = TRUE,
panel.labels = c("AM (06:00 - 10:00)", "PM (16:00 - 20:00)"),
legend.outside = T, title = "N. trips/year") +
tm_facets(ncol = 2)
m
tmap_save(tm = m, "figures/map-am-pm-peaks.png")
```
Get OAC data (not evaluated)
```{r, eval=FALSE}
devtools::install_github("robinlovelace/ukboundaries")
# download.file("https://data.cdrc.ac.uk/dataset/68771b14-72aa-4ad7-99f3-0b8d1124cb1b/resource/8fff55da-6235-459c-b66d-017577b060d3/download/output-area-classification.zip", "output-area-classification.zip") # fails
oac = sf::read_sf("Output Area Classification/Shapefiles/2011_OAC.shp")
u_oas_cents = "https://opendata.arcgis.com/datasets/ba64f679c85f4563bfff7fad79ae57b1_0.zip?outSR=%7B%22wkid%22%3A27700%2C%22latestWkid%22%3A27700%7D"
u_oas_cents = ukboundaries::duraz(u_oas_cents)
oas_lnd = u_oas_cents %>% st_transform(4326)
oas_lnd = oas_lnd[stations_region, ] # works
oac_lnd = oac %>% filter(OA_SA %in% oas_lnd$oa11cd)
plot(oac_lnd["SPRGRP"])
# IMD data (at LSOA level)
# imd = readr::read_csv("https://opendata.arcgis.com/datasets/da3b33dd44d94f48a9628a3391957505_0.csv") # seems to be ranks...
# imd = readr::read_csv("https://data.cdrc.ac.uk/dataset/4d3a8738-38af-401c-8070-6be5d85b2f5e/resource/a4230484-9104-4903-b754-0d07a1278862/download/imd2015eng.csv") # navigate to that URL and manually download...
imd = readr::read_csv("imd2015eng.csv")
lsoa = ukboundaries::lsoa2011_simple
lsoa_lnd = lsoa[oas_lnd, ]
imd = imd %>%
rename(lsoa11cd = `LSOA code (2011)` )
lsoa_lnd = inner_join(lsoa_lnd, imd)
plot(lsoa_lnd["Index of Multiple Deprivation (IMD) Score"])
sf::write_sf(lsoa_lnd, "lsoa_bikeshare.geojson")
oas_lnd_imd = st_join(oas_lnd, lsoa_lnd)
plot(oas_lnd_imd["Index of Multiple Deprivation (IMD) Score"])
sf::write_sf(oas_lnd_imd, "oas_bikeshare_imd.geojson")
names(oas_lnd_imd) = snakecase::to_snake_case(names(oas_lnd_imd))
names(oas_lnd)
plot(oas_lnd_imd[7:20])
# sf::write_sf(oas_lnd_imd, "oas_bikeshare_imd_tidynames.geojson")
# piggyback::pb_upload("oas_bikeshare_imd_tidynames.geojson")
# piggyback::pb_upload("oas_bikeshare_imd.geojson")
# piggyback::pb_upload("lsoa_bikeshare.geojson")
```
# Classify docking stations
Based on number of nearby residential zones and IMD/income scores.
Initial method: by distance to docking stations (flaw with 300m: misses reality of walking that far, try 200m).
```{r}
oas = read_sf("oas_bikeshare_imd_tidynames.geojson")
if(!file.exists("sape21dt10hmid2018east.zip")) {
u_oas_pop = "https://www.ons.gov.uk/file?uri=%2fpeoplepopulationandcommunity%2fpopulationandmigration%2fpopulationestimates%2fdatasets%2fcensusoutputareaestimatesintheeastregionofengland%2fmid2018sape21dt10h/sape21dt10hmid2018east.zip"
download.file(u_oas_pop, "sape21dt10hmid2018east.zip")
unzip("sape21dt10hmid2018east.zip")
}
# mid 2018 estimates: https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates/datasets/censusoutputareaestimatesintheeastregionofengland
# oas_pop = readxl::read_excel("SAPE21DT10h-mid-2018-coa-unformatted-syoa-estimates-east.xlsx", sheet = 4)
# oas_name = as.character(oas_pop[4, ])
# oas_pop = oas_pop %>% slice(5:nrow(oas_pop))
# names(oas_pop) = oas_name
# oas_pop_16plus = oas_pop %>% mutate(Population = rowSums(select(., 20:(ncol(oas_pop) - 1)))) %>%
# select(OA11CD, Population)
# see https://data.london.gov.uk/dataset/statistical-gis-boundary-files-london
if(!file.exists("statistical-boundaries-london.zip")) {
download.file("https://data.london.gov.uk/download/statistical-gis-boundary-files-london/9ba8c833-6370-4b11-abdc-314aa020d5e0/statistical-gis-boundaries-london.zip", "statistical-boundaries-london.zip")
unzip("statistical-boundaries-london.zip")
}
oaps_all_london = sf::read_sf("statistical-gis-boundaries-london/ESRI/OA_2011_London_gen_MHW.shp")
plot(oaps_all_london$geometry)
oaps = oaps_all_london %>% st_transform(4326)
oaps = oaps[stations_region, ]
plot(oaps)
oaps = oaps %>% select(OA11CD, Populaton2011 = USUALRES)
plot(oaps)
stations = sf::read_sf("stations_clean_am_pm.geojson")
plot(stations["Income.decile"])
stations_150m_buffer = stations %>%
st_transform(27700) %>%
st_buffer(150) %>%
st_transform(4326)
mapview::mapview(stations_150m_buffer)
# aim: get to trips per person within 200m per docking station
# geofrabric method fails - see https://github.com/ITSLeeds/geofabric/issues/16
# lnd_osm = ukboundaries::duraz("http://download.geofabrik.de/europe/great-britain/england/greater-london-latest-free.shp.zip")
# lnd_osm = sf::read_sf("gis_osm_buildings_a_free_1.shp")
# lnd_osm_stations = lnd_osm[stations_region, ]
# saveRDS(lnd_osm_stations, "lnd_osm_stations.Rds")
# piggyback::pb_upload("lnd_osm_stations.Rds")
lnd_osm_stations = readRDS("lnd_osm_stations.Rds")
# plot(lnd_osm_stations$geometry) # slow to plot
lnd_osm_resi = lnd_osm_stations %>%
filter(str_detect(string = type, pattern = "house|resi|apartments|terrace|flats|council_flats") |
is.na(type))
plot(lnd_osm_resi$geometry, col = "red", border = "red", add = TRUE)
# idea: use building centroids and calculate n. people per building
lnd_osm_centroids = st_centroid(lnd_osm_resi)
table(lnd_osm_centroids$OA11CD) %>% head()
lnd_osm_centroids = st_join(lnd_osm_centroids, oaps)
# get number of buildings per zones
lnd_osm_centroids_with_count = lnd_osm_centroids %>% dplyr::add_count(OA11CD)
lnd_osm_centroids$n = lnd_osm_centroids_with_count$n
mapview::mapview(lnd_osm_centroids %>% filter(n == max(n)))
lnd_osm_centroids = lnd_osm_centroids %>% mutate(pop_per_building = Populaton2011 / n)
lnd_osm_near_stations = lnd_osm_centroids[stations_150m_buffer, ]
mapview::mapview(stations_150m_buffer) + mapview::mapview(lnd_osm_near_stations)
stations_pop = aggregate(lnd_osm_near_stations["pop_per_building"], stations_150m_buffer, sum)
stations_pop = stations_pop %>% rename(Population = pop_per_building)
stations_pop
stations = st_join(stations, stations_pop)
sum(stations_pop$Population, na.rm = TRUE)
sum(lnd_osm_centroids$pop_per_building, na.rm = TRUE)
plot(stations_pop)
# mapview::mapview(stations_pop) + mapview::mapview(lnd_osm_centroids)
saveRDS(stations, "stations_pop.Rds")
piggyback::pb_upload("stations_pop.Rds")
# write_sf(stations, "stations_pop.geojson", delete_dsn = TRUE)
# plot result
m = tm_shape(oaps) + tm_fill("Populaton2011", palette = "Blues", title = "Population") +
tm_shape(lnd_osm_centroids) + tm_dots(col = "green", alpha = 0.2) +
tm_shape(lnd_osm_near_stations) + tm_dots(col = "red", alpha = 0.6) +
tm_shape(stations) + tm_dots(size = 0.07)
m
tmap_save(m, "figures/bikeshare-resi-buildings.png")
```
## Now use the population estimates to show AM peak usage (per year?)
```{r}
stations = readRDS("stations_pop.Rds")
names(stations)
stations = stations %>% rename(income_decile = Income.decile)
summary(stations$income_decile)
stations$trips_per_person_yr = stations$trips_per_year / stations$Population / stations$years_in_operation
stations$trips_per_person_yr_am = stations$trips_per_year_am / stations$Population
stations$trips_per_person_yr_pm = stations$trips_per_year_pm / stations$Population
summary(stations$trips_per_person_yr)
stations$trips_per_person_yr[stations$trips_per_person_yr > 5]
unique(stations$income_decile)
stations$income_decile = formatC(x = stations$income_decile, width = 2, flag = "0")
class(stations$income_decile) = "character"
# previously was ggbeeswarm::geom_beeswarm, replaced by geom_violin
theme_set(theme_minimal())
g1 = ggplot(stations %>% filter(! income_decile == "NA")) + geom_boxplot(aes(income_decile, trips_per_person_yr), draw_quantiles = c(0.25, 0.5, 0.75)) + scale_y_continuous(limits = c(0.1, 20)) + ylab("Number of trips per local resident")
g1
g2 = ggplot(stations %>% filter(! income_decile == "NA")) + geom_boxplot(aes(income_decile, trips_per_person_yr_am), draw_quantiles = c(0.25, 0.5, 0.75)) + scale_y_continuous(limits = c(0.1, 20)) + ylab("Number of trips per local resident (AM)") + xlab("Income decile")
g2
g3 = ggplot(stations %>% filter(! income_decile == "NA")) + geom_boxplot(aes(income_decile, trips_per_person_yr_pm), draw_quantiles = c(0.25, 0.5, 0.75)) + scale_y_continuous(limits = c(0.1, 20)) + ylab("Number of trips per local resident (PM)") + xlab("Income decile")
g3
library(patchwork)
p = g2 + g3 + plot_layout(ncol = 1)
p
ggsave("figures/income-decile-am-pm-boxplot.png", p, width = 9, height = 7)
magick::image_read("figures/income-decile-am-pm-boxplot.png")
# ggplot(stations %>% filter(!is.na(income_decile))) + geom_boxplot(aes(group = income_decile, y = trips_per_person_yr)) + scale_x_discrete(labels = 1:10)
summary(stations$income_decile)
g = ggplot(stations) + geom_point(aes(Population, total_n_trips_start))
g
cor(stations$Population, stations$total_n_trips_start, use = "complete.obs")
cor(stations$Population, stations$trips_per_year_am, use = "complete.obs")
# postive correlation, expected
quantile(x = stations$trips_per_person_yr, probs = c(0.1, 0.5, 0.7, 0.9), na.rm = T)
# check deciles - are there really that many deprived oas in London?
# file.remove("stations_imd_pop.geojson")
write_sf(stations, "stations_imd_pop.geojson", delete_dsn = TRUE)
piggyback::pb_upload("stations_imd_pop.geojson")
# hist(stations$imd)
```
## Exploration of residential zones
```{r, eval=FALSE}
tmap_mode("view")
tm_shape(stations) + tm_dots(col = "n_oas", size = 0.1)
tmap_mode("plot")
qtm(stations %>% filter(n_oas <= 2))
stations_residential = stations %>% filter(trips_per_person_yr_am > 1, n_oas >= 1)
tm_shape(stations_residential) + tm_dots()
stations_residential %>%
filter(trips_per_person_yr_am > 5) %>%
tm_shape() + tm_dots() # excludes rail station
nrow(stations_residential)
summary(stations_residential$trips_per_person_yr)
table(stations_residential$n_oas)
summary(stations_residential$Population)
# 0.16 with outliers removed
cor(stations_residential$Population, stations_residential$total_n_trips_start, use = "complete.obs") # positive cor
# write_sf(stations_residential, "stations_residential.geojson")
# piggyback::pb_upload("stations_residential.geojson")
tm_shape(stations) + tm_dots("grey") +
tm_shape(stations_residential) + tm_dots()
sf::write_sf(stations_residential, "stations_residential.geojson", delete_dsn = TRUE)
piggyback::pb_upload("stations_residential.geojson")
```
Subsetting only OD pairs with origins in the 'residential stations'.
```{r, eval=FALSE}
trips_df$id = as.character(trips_df$id)
stations_residential$id
unique(trips_df$start_station_id)
trips_df_originating_in_residential_zones = trips_df$start_station_id %in% stations_residential$id
summary(trips_df_originating_in_residential_zones)
sum(trips_df_originating_in_residential_zones) / nrow(trips_df) # 40% trips remain
trips_resi = trips_df %>%
filter(trips_df_originating_in_residential_zones)
summary(trips_resi$start_station_id %in% stations_residential$id)
summary(trips_resi$end_station_id %in% stations_residential$id)
fst::write.fst(trips_resi, "london_bike_hire_cleaned_residential.fst")
piggyback::pb_upload("london_bike_hire_cleaned_residential.fst")
# vroom::vroom_write(trips_resi, "london_bike_hire_cleaned_residential.csv.gz")
# piggyback::pb_upload("london_bike_hire_cleaned_residential.csv.gz")
```
# Quantifying inequality in provision, compared with randomly sampled points
How does the distribution of IMD scores associated with the stations compare with IMD scores associated with randomly sampled points?
```{r, eval=FALSE}
sdf = read_sf("sdf.geojson")
set.seed(1985)
random_points = st_sample(x = sdf[4, ], size = nrow(stations))
random_points_oas = aggregate(oas["index_of_multiple_deprivation_imd_score"], random_points, mean, join = j)
random_points_oas = aggregate(oas["index_of_multiple_deprivation_imd_score"], random_points, mean, join = j)
hist(random_points_oas$index_of_multiple_deprivation_imd_score)
hist(stations$imd)
```
```{r cleaning-old, eval=FALSE, echo=FALSE}
# from CDRC
piggyback::pb_download("stations_ucl_yearly.geojson")
piggyback::pb_download("stations2019.geojson")
plot(stations, add = TRUE)
length(unique(stations$ucl_id))
# cleaning stations data
stations = stations %>% ungroup() %>% rename(id = ucl_id) %>% mutate(id = as.character(id))
table(stations$year) # trust the 2012, 2014, 2015, 2016 ones
plot(stations$initial_size, stations$curr_size) # interesting contrast
# ids in bikedata but not ucl
(ids_in_bikedata_not_in_clean_ucl_data = stations_bikedata$id[! stations_bikedata$id %in% stations$id])
(ids_in_bikedata_not_in_ucl_data = stations_bikedata$id[! stations_bikedata$id %in% stations$id])
(ids_in_ucl_data_not_in_bikedata = stations$id[! stations$id %in% stations_bikedata$id])
```
```{r, eval=FALSE, echo=FALSE}
# out-takes
# old method
summary(st_area(oas))
plot(lsoa_lnd$geometry)
(sum(st_area(lsoa_lnd)) / nrow(oas))^0.5 # on average around 200 apart
sel_200 = st_is_within_distance(stations, oas, 200)
summary(lengths(sel_200)) # catches on average 5 areas
j = function(x, y) st_is_within_distance(x, y, dist = 200)
# stations_oas = aggregate(oas["index_of_multiple_deprivation_imd_score"], stations, weighted.mean, join = j, w = oas$total_population_mid_2012_excluding_prisoners) # fails - w not of right length
stations_oas = aggregate(oas["index_of_multiple_deprivation_imd_score"], stations, mean, join = j)
stations$imd = stations_oas$index_of_multiple_deprivation_imd_score
plot(stations["imd"])
summary(stations$imd)
plot(stations_oas) # they are the same
stations_pop = aggregate(oas["population_aged_16_59_mid_2012_excluding_prisoners"], stations, sum, join = j)
stations_n_oas = aggregate(oas["population_aged_16_59_mid_2012_excluding_prisoners"], stations, length, join = j)
stations_median_decile = aggregate(oas["index_of_multiple_deprivation_imd_decile_where_1_is_most_deprived_10_of_lso_as"], stations, median, join = j)
stations_median_income = aggregate(oas["income_decile_where_1_is_most_deprived_10_of_lso_as"], stations, median, join = j)
names(stations_n_oas)[1] = "n_oas_200m"
plot(stations_pop)
plot(stations_n_oas)
hist(stations_n_oas$n_oas_200m)
# <!-- Let's plot what just happened. Let's take station 1: -->
oas1 = oas[stations[1, ], , op = st_is_within_distance, dist = 200]
mapview::mapview(oas1) + mapview::mapview(stations[1, ])
# Note: The distance for 300m buffer looked far for the outer stations. Note: try with lower distance in future.
# For now, let's add the new aggregated values and save the result:
# stations$pop_centroid_within_200m = stations_pop$population_aged_16_59_mid_2012_excluding_prisoners
stations$n_oas = stations_n_oas$n_oas_200m
stations$imd_decile_median_where_1_is_most_deprived = stations_median_decile$index_of_multiple_deprivation_imd_decile_where_1_is_most_deprived_10_of_lso_as %>% round()
plot(stations$imd_decile_median_where_1_is_most_deprived)
stations$income_decile = stations_median_income$income_decile_where_1_is_most_deprived_10_of_lso_as %>% round()
```