-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathnyc.R
278 lines (205 loc) · 8.83 KB
/
nyc.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
library(tidyverse)
inp1<-read.table("train.csv",sep=",",header=T,stringsAsFactors = F)
inp<-as.tibble(inp1)
#-- geo coordinates
trip.start <- inp %>% dplyr::select(id, pickup_longitude,pickup_latitude)
trip.end <- inp %>% dplyr::select(id, dropoff_longitude,dropoff_latitude)
library(sp)
library(geojsonio)
library(spatialEco)
coordinates(trip.start) <- ~pickup_longitude+pickup_latitude
coordinates(trip.end)<- ~dropoff_longitude+dropoff_latitude
#----------------------------------------------------------------
# reverse geocoding: borough
#----------------------------------------------------------------
boroughs<-geojson_read("Borough Boundaries.geojson",what="sp")
trip.start@proj4string<-boroughs@proj4string
trip.end@proj4string<-boroughs@proj4string
start.joined <- point.in.poly(trip.start, boroughs)
end.joined <- point.in.poly(trip.end, boroughs)
start<-data.frame(id=start.joined$id,start_boro_name=start.joined$boro_name,stringsAsFactors = F)
end<-data.frame(id=end.joined$id,end_boro_name=end.joined$boro_name,stringsAsFactors = F)
inp1<-inp %>% left_join(start) %>% left_join(end)
inp<-inp1
#----------------------------------------------------------------
# reverse geocoding: neighbourhood
#----------------------------------------------------------------
nbrh<-geojson_read("Neighborhood Tabulation Areas.geojson",what="sp")
trip.start@proj4string<-nbrh@proj4string
trip.end@proj4string<-nbrh@proj4string
start.joined <- point.in.poly(trip.start, nbrh)
end.joined <- point.in.poly(trip.end, nbrh)
start<-data.frame(
id=start.joined$id,
start_county=start.joined$county_fips,
start_ntaname=start.joined$ntaname,
stringsAsFactors = F)
end<-data.frame(
id=end.joined$id,
end_county=end.joined$county_fips,
end_ntaname=end.joined$ntaname,
stringsAsFactors = F)
inp<-inp %>% left_join(start) %>% left_join(end)
#----------------------------------------------------------------
# reverse geocoding: zipcode
#----------------------------------------------------------------
postal <- geojson_read("nyc-zip-code-tabulation-areas-polygons.geojson",what="sp")
trip.start@proj4string<-postal@proj4string
trip.end@proj4string<-postal@proj4string
start.joined <- point.in.poly(trip.start, postal)
end.joined <- point.in.poly(trip.end, postal)
start<-data.frame(id=start.joined$id,start_postal=start.joined$postalCode,start_poname=start.joined$PO_NAME,stringsAsFactors = F)
end<-data.frame(id=end.joined$id,end_postal=end.joined$postalCode,end_poname=end.joined$PO_NAME,stringsAsFactors = F)
inp<-inp %>% left_join(start) %>% left_join(end)
#-----------------------
# more massaging
#-----------------------
inp<- inp %>% mutate(start_day=substr(pickup_datetime,1,10),start_time=substr(pickup_datetime,12,19) )
#----------------------
# add weather
#----------------------
weather <- read.table("weather.tsv",sep="\t",header=T,stringsAsFactors = F)
w.full <- weather %>%
filter(boro_name=="Queens") %>%
mutate(boro_name="Bronx") %>%
rbind(weather) %>%
rename(start_boro_name = boro_name)
inp <- inp %>% left_join(w.full)
#---------------------
# add sunrise/sunset
#---------------------
sunrise <- read.table("sunrise.txt",sep="\t",header=T,stringsAsFactors = F)
inp.tmp <- inp %>% left_join(sunrise)
inp<-inp.tmp
inp.tmp <- inp %>%
mutate(start_day=ymd(start_day))
inp<-inp.tmp
inp <- inp %>%
mutate(before_sunrise=ifelse(start_time<sunrise,1,0)) %>%
mutate(after_sunset=ifelse(start_time>sunset,1,0)) %>%
mutate(is_dark=sign(before_sunrise+after_sunset)) %>%
mutate(before_twilight=ifelse(start_time<twilight_start,1,0)) %>%
mutate(after_twilight=ifelse(start_time>twilight_end,1,0)) %>%
mutate(outside_twilight=sign(before_twilight+after_twilight))
#---------------------
# add calendar events
#---------------------
inp <- inp %>% mutate(day_of_week = wday(start_day,label=T))
inp$holiday<-0
#Friday January 01 New Years Day
inp[inp$start_day=="2016-01-01",]$holiday<-1
#Monday January 18 Martin Luther King Jr. Day 3rd Monday in January
inp[inp$start_day=="2016-01-18",]$holiday<-1
#Friday February 12 Lincoln's Birthday Connecticut, Illinois, Missouri, New Jersey, New York.
#Monday February 15 Presidents' Day 3rd Monday in February. Not all states
inp[inp$start_day=="2016-02-15",]$holiday<-1
#Monday May 30 Memorial Day Last Monday in May
inp[inp$start_day=="2016-05-30",]$holiday<-1
#---------------------
# add accidents
#---------------------
accidents<-read.table("Austin_SES_Demo_NYC_Traffic.tsv",sep="\t", header=T, stringsAsFactors = F,fill=T)
library(stringr)
accidents <- accidents %>% mutate(TIME=str_pad(TIME,5,0,side = "left"))
accidents <- accidents %>% mutate(DATE=mdy(DATE))
acc_id<-(1:nrow(accidents))
accidents$acc_id<-acc_id
accidents <-accidents %>% filter(str_length(str_trim(BOROUGH))>0)
accidents <-accidents %>% mutate(boro_name=str_to_title(BOROUGH))
tmp.inp <- inp %>% dplyr::select(id,pickup_longitude,pickup_latitude,start_boro_name,start_day,start_time) %>%
mutate(start_day=ymd(start_day))
hrs<-as.numeric(substr(tmp.inp$start_time,1,2))*60
mins<-as.numeric(substr(tmp.inp$start_time,4,5))
tmp.time<-hrs+mins
tmp.inp$time<-tmp.time
tmp.time<-as.numeric(substr(accidents$TIME,1,2))*60+as.numeric(substr(accidents$TIME,4,5))
accidents$acc_time<-tmp.time
inp$start_day<-ymd(inp$start_day)
accidents.start <- accidents %>% dplyr::select(boro_name,DATE,LATITUDE,LONGITUDE,acc_id,acc_time) %>%
rename(start_boro_name=boro_name,start_day=DATE) %>%
inner_join(tmp.inp)
accidents.start$td <- abs(accidents.start$acc_time-accidents.start$time)
accidents.start <- accidents.start %>% filter(td<=10)
write.table(accidents.start,"accidents_start.txt",sep="|",row.names=F)
accidents.start$dist<-0
dists<-accidents.start$dist
pb <- txtProgressBar(min = 0, max = nrow(accidents.start), initial = 0)
for(i in 1:nrow(accidents.start)){
r<-accidents.start[i,]
dists[i]<-dist_fun(r$LONGITUDE,r$LATITUDE,r$pickup_longitude,r$pickup_latitude)
setTxtProgressBar(pb, i)
}
accidents.start <- accidents.start %>% mutate(TIME=hm(TIME),start_time=hms(start_time))
#-- add time difference: accidents withing 10 minutes of trip start
#-------------------------------
# add non NYC data
#-------------------------------
states<-geojson_read("gz_2010_us_040_00_5m.json",what="sp")
trip.start@proj4string<-states@proj4string
trip.end@proj4string<-states@proj4string
start.joined <- point.in.poly(trip.start, states)
end.joined <- point.in.poly(trip.end, states)
start<-data.frame(id=start.joined$id,start_state=start.joined$NAME,stringsAsFactors = F)
end<-data.frame(id=end.joined$id,end_state=end.joined$NAME,stringsAsFactors = F)
inp1<-inp %>% left_join(start) %>% left_join(end)
inp<-inp1
#-------- cleanup
inp[is.na(inp$start_boro_name),]$start_boro_name <- "OutsideNYC"
inp$start_ntaname<-as.character(inp$start_ntaname)
inp[is.na(inp$start_ntaname),]$start_ntaname <- "OutsideNYC"
inp$start_poname <- as.character(inp$start_poname)
inp[is.na(inp$start_poname),]$start_poname <- "OutsideNYC"
inp$start_postal<-as.character(inp$start_postal)
inp[is.na(inp$start_postal),]$start_postal<-"00000"
#--
inp$end_boro_name<-as.character(inp$end_boro_name)
inp[is.na(inp$end_boro_name),]$end_boro_name <- "OutsideNYC"
inp$end_ntaname<-as.character(inp$end_ntaname)
inp[is.na(inp$end_ntaname),]$end_ntaname <- "OutsideNYC"
inp$end_poname <- as.character(inp$end_poname)
inp[is.na(inp$end_poname),]$end_poname <- "OutsideNYC"
inp$end_postal<-as.character(inp$end_postal)
inp[is.na(inp$end_postal),]$end_postal<-"00000"
inp$start_county<-NULL
inp$end_county<-NULL
inp$start_state<-as.character(inp$start_state)
inp$end_state<-as.character(inp$end_state)
inp$outside<-0
inp[inp$start_boro_name=="OutsideNYC",]$outside<-1
inp[inp$end_boro_name=="OutsideNYC",]$outside<-1
#-------------------------------
# add start/destination distance
#-------------------------------
inp <- inp %>%
mutate(p2p=paste0(start_boro_name,'-',end_boro_name)) %>%
mutate(same_boro = ifelse(start_boro_name==end_boro_name,1,0))
inp[inp$outside==1,]$same_boro<-0
library(geosphere)
dist_fun<-function(long1,lat1,long2,lat2){
distm (c(long1, lat1), c(long2, lat2), fun = distGeo)
}
dists<-inp$id
for(i in 1:nrow(inp)){
r<-inp[i,]
dists[i] <- dist_fun(r$pickup_longitude,r$pickup_latitude,r$dropoff_longitude,r$dropoff_latitude)
}
#--------------------- outliers
#-- duration outlier
inp$outlier<-0
inp[inp$trip_duration>mean(inp$trip_duration)+3*sd(inp$trip_duration),]$outlier<-1
inp<-inp %>% rename(time_outlier=outlier)
#-- distance outlier
inp$dist_outlier<-0
inp[inp$dist>mean(inp$dist)+3*sd(inp$dist),]$dist_outlier<-1
#1 572 records
inp[inp$dist>30000,]$dist_outlier<-1
#------------------------------
# exploration
#------------------------------
inp$dist<-as.numeric(inp$dist)
summary(inp$dist)
#------------------------------
# save data
#------------------------------
write.table(inp,"train_processed.txt",sep="|",row.names=F)
inp <- read.table("train_processed.txt",sep="|",stringsAsFactors =F,header=T)