-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathKernelDensityAnalysis.R
241 lines (184 loc) · 8.79 KB
/
KernelDensityAnalysis.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
setwd("~/Library/Mobile Documents/com~apple~CloudDocs/OzoneHeatJoint")
source("ozoneheatwavedefs.R")
dataPrep <- function(){
print("Loading Datasets")
all <- read.csv("Whole\ CA\ Respi.csv", header = T)
summary(all$date)
#popZCTA <- read.csv("2010Census_DemoProfile.csv", header = TRUE)
all$date2 <- as.Date(all$date, format = "%d%b%Y")
all.days <- all[which(all$date2 > as.Date("2004-01-01")),]
# ozone <- read.csv("ca_pollutants_90t17.csv", header = T)
# ozone$date2 <- as.Date(ozone$date, format = "%d%b%Y")
# ozone <- ozone[which(ozone$date2 > as.Date("2004-01-01")),]
ozone = read.csv("Ozone_2004_13.csv")
ozone$date2 <- as.Date(ozone$date, format = "%Y-%m-%d")
#ozone <- ozone[which(ozone$date2 > as.Date("2004-01-01")),]
ozone$max_o3 = ozone$ozone_idw
ozone$zipcode = ozone$zcta5
temps <- read.csv("ca_zip_temp_04t13.csv", header = TRUE)
temps$date2 <- as.Date(temps$date, format = "%Y-%m-%d")
temps2 <- temps[which(temps$date2 > as.Date("2004-01-01")),]
temps2$zipcode <- temps2$zipnum
temps2$zipnum = NULL
library(tidyr)
library(dplyr)
library(ggplot2)
library(lme4)
all.days <- inner_join(all.days, temps2, by = c("patzip" = "zipcode", "date2"))
#summary(ozone$date)
#data(zipcode)
#zipcode$zip = readr::parse_number(zipcode$zip)
testset <- inner_join(all.days, ozone, by = c("patzip" = "zipcode", "date2"))
testset$weekday <- weekdays(testset$date2)
testset$month <- format(testset$date2,"%B")
newtest <- subset(testset, testset$month == "May" | testset$month == "June" | testset$month == "July" | testset$month == "August" | testset$month == "September")
newtest$month <- as.factor(newtest$month)
newtest$weekday <- as.factor(newtest$weekday)
newtest$year <- format(newtest$date2, "%Y")
newtest <- within(newtest, month <- relevel(month, ref = "May"))
newtest <- within(newtest, weekday <- relevel(weekday, ref = "Monday"))
library(dplyr)
library(Hmisc)
newtest <-
newtest %>%
group_by(patzip) %>%
mutate(l1resp = dplyr::lead(resp_disease, n = 1, default = NA))
newtest <-
newtest %>%
group_by(patzip) %>%
mutate(l2resp = dplyr::lead(resp_disease, n = 2, default = NA))
newtest$resp_Avglead2 = rowMedians(as.matrix(newtest[, c("l1resp", "l2resp")]))
newtest$resp_Avglead2 = ifelse(is.na(newtest$resp_Avglead2), newtest$resp_disease, newtest$resp_Avglead2)
#data(zipcode)
#zipcode$zip = readr::parse_number(zipcode$zip)
#newtest <- left_join(newtest, zipcode, by = c("patzip" = "zip"))
return(newtest)
}
newtest <- dataPrep()
newtest <- ozoneDefs(newtest)
newtest <- heatwaveDefs(newtest)
library(dplyr)
data.frame(table(newtest$patzip))
#1530 days for each zip code
test_missingness=newtest %>%
group_by(patzip) %>%
summarise_all(~sum(is.na(max_o3))) %>%
transmute(patzip, sumNA = rowSums(.[-1]))
# jointOHWave <- function(x, ozoneDef, heatwaveDef){
# x$jointOH <- ifelse(ozoneDef == 1 & heatwaveDef == 1, 1, ifelse(heatwaveDef == 1 | ozoneDef == 1, -1, 0))
# x$onlyHW <- ifelse(ozoneDef == 0 & heatwaveDef == 1, 1, 0)
# x$onlyOz <- ifelse(ozoneDef == 1 & heatwaveDef == 0, 1, 0)
# return(x)
# }
# newtest <- jointOHWave(newtest, newtest$o75wave, newtest$hw99_1)
#newtest$max_o3 <- 1000*newtest$max_o3
missingness_byzip = numeric()
j = 1
for(i in unique(newtest$patzip)){
temp = newtest[which(newtest$patzip == i),]
missingness_byzip[j]=sum(is.na(temp$max_o3))/nrow(temp)
j = j + 1
}
#h = missingness_byzip[which(missingness_byzip != 1)]
# Inverse distance in time metric for creating matching
# Take all the data from the zip code and same year with exposed days
# Use date subtraction in R to compute "distance" between two dates (to each case day)
#1/distance is your weight
# Take the hospitalizations on those dates and create a weighted average, ie you multiply the hospitalizations
#by the weight and then divide by the sum of the weights
KernelDistanceMatching <- function(case, control_dataset){
weightedAvgDiff = numeric(0)
print(nrow(case))
for(i in 1:nrow(case)){
if(i %% 1000 == 0) print(i)
controls <- control_dataset[which(control_dataset$patzip == case$patzip[i] & control_dataset$year == case$year[i]),]
controls$distance = as.numeric(abs(case$date2[i] - controls$date2), units = "days")
controls$weight = 1/controls$distance
weightedAvgDiff[i] = case$resp_Avglead2[i] - sum(controls$resp_Avglead2*controls$weight)/sum(controls$weight)
}
return(weightedAvgDiff)
}
MatchProcedure <- function(x){
dataMatch <- x
dataMatch$month <- as.numeric(dataMatch$month)
Xformatch <- dataMatch[,c("patzip","year","date2","hw95_1", "o75wave", "resp_disease", "resp_Avglead2", "latitude", "longitude")]
#Get Contrast for Joint Days
XJoint <- Xformatch[which(Xformatch$hw95_1 == 1 & Xformatch$o75wave == 1),]
print(nrow(XJoint))
zipcodesJoint <- unique(XJoint$patzip)
XNon <- Xformatch[which(Xformatch$hw95_1 == 0 & Xformatch$o75wave == 0),]
zipcodesTotal <- unique(XNon$patzip)
XJoint$diffMean <- KernelDistanceMatching(XJoint, XNon)
#Get contrast for heat wave only days
XHeat <- Xformatch[which(Xformatch$hw95_1 == 1 & Xformatch$o75wave == 0),]
XHeat$diffMean <- KernelDistanceMatching(XHeat, XNon)
#Get contrast for ozone wave only days
XOzone <- Xformatch[which(Xformatch$hw95_1 == 0 & Xformatch$o75wave == 1),]
XOzone$diffMean <- KernelDistanceMatching(XOzone,XNon)
return(list(Joint = XJoint, HW = XHeat, OZ = XOzone))
}
matched = MatchProcedure(newtest)
#write.csv(matched$Joint, "joint975_2.csv")
#write.csv(matched$HW, "hwonly975_2.csv")
#write.csv(matched$OZ, "ozonly975_2.csv")
matched$Joint$RR <- matched$Joint$resp_disease/(matched$Joint$resp_disease - matched$Joint$diffMean)
matched$HW$RR <- matched$HW$resp_disease/(matched$HW$resp_disease - matched$HW$diffMean)
matched$OZ$RR <- matched$OZ$resp_disease/(matched$OZ$resp_disease - matched$OZ$diffMean)
# matched$Joint$RR <- matched$Joint$resp_Avglead2/(matched$Joint$resp_Avglead2 - matched$Joint$diffMean)
# matched$HW$RR <- matched$HW$resp_Avglead2/(matched$HW$resp_Avglead2 - matched$HW$diffMean)
# matched$OZ$RR <- matched$OZ$resp_Avglead2/(matched$OZ$resp_Avglead2 - matched$OZ$diffMean)
matched$Joint <- matched$Joint[which(is.finite(matched$Joint$RR)),]
matched$HW <- matched$HW[which(is.finite(matched$HW$RR)),]
matched$OZ <- matched$OZ[which(is.finite(matched$OZ$RR)),]
RRjoint = aggregate(matched$Joint$RR, list(matched$Joint$patzip), mean)
RRjoint$jointRR = RRjoint$x
RRjoint$x = NULL
RRhw =aggregate(matched$HW$RR, list(matched$HW$patzip), mean)
RRhw$hwRR = RRhw$x
RRhw$x = NULL
RRoz = aggregate(matched$OZ$RR, list(matched$OZ$patzip), mean)
RRoz$ozRR = RRoz$x
RRoz$x = NULL
allRR = inner_join(RRjoint, RRhw, by = c("Group.1"))
allRR = inner_join(allRR, RRoz, by = c("Group.1"))
allRR$RERI = allRR$jointRR - allRR$hwRR - allRR$ozRR + 1
allRR$RERI <- if_else(allRR$RERI < -3,-3,allRR$RERI)
allRR$RERI <- if_else(allRR$RERI > 3, 3, allRR$RERI)
allRR$ZCTA5CE10 <- allRR$Group.1
# write.csv(allRR, "allRR_975_2.csv")
# library(sf)
missing =data.frame(cbind(unique(newtest$patzip), missingness_byzip))
names(missing) <- c("ZCTA5CE10", "missing")
shape <- st_read(dsn = "~/Downloads/ShapefileZCTA/cb_2016_us_zcta510_500k.shp")
m <- merge(shape, missing, by = ("ZCTA5CE10"))
#
# allRR$hwRR = if_else(allRR$hwRR > 6, 6, allRR$hwRR)
# allRR$ozRR = if_else(allRR$ozRR>6, 6, allRR$ozRR)
# nonEnviro = as.data.frame(matrix(nrow = length(setdiff(unique(newtest$patzip), RRoz$Group.1)),ncol = ncol(allRR)))
# names(nonEnviro) = names(allRR)
# nonEnviro$ZCTA5CE10 = setdiff(unique(newtest$patzip), RRoz$Group.1)
#
# allRR = rbind(nonEnviro,allRR)
# m <- merge(shape, allRR, by = ("ZCTA5CE10"))
# par(mfrow = c(1,2))
MainStates <- map_data("state")
NewStates <- filter(MainStates,region == "california" )
# #hw RR
plot1 <- ggplot() +
geom_sf(data = m, aes(fill = missing), colour = NA) + geom_polygon( data=NewStates, aes(x=long, y=lat, group=group),
colour="black", fill = NA)
#
plot1 + scale_fill_gradient2(low = "white",high="red", limits = c(0,0.99)) + theme(legend.key.size = unit(1, "cm")) +labs(fill = "missingness") + ggtitle("Missingness in O3 estimates") + theme(panel.grid = element_blank())
# #oz RR
# plot2 <- ggplot() +
# geom_sf(data = m, aes(fill = ozRR), colour = NA) + geom_polygon( data=NewStates, aes(x=long, y=lat, group=group),
# colour="black", fill = NA)
#
# plot2 <- plot2 + scale_fill_gradient2(low = "blue",mid = "white",
# high="red", midpoint = 1, limits = c(0,6)) + theme(legend.key.size = unit(1, "cm")) +
# labs(fill = "RR") + ggtitle("Ozone relative risk") + theme(panel.grid = element_blank())
#
# require(gridExtra)
# pdf("RRplotgrid.pdf")
# grid.arrange(plot1, plot2, ncol = 2)
# dev.off()