-
Notifications
You must be signed in to change notification settings - Fork 0
/
traffic mort - RSA.R
211 lines (151 loc) · 8.5 KB
/
traffic mort - RSA.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
require(data.table)
require(reshape2)
require(dplyr)
require(stringr)
require(ggplot2)
require(plyr)
### Death data - WHO - Cause of Death
# these files are available from http://www.who.int/healthinfo/statistics/mortality_rawdata/en/),
file1 = fread("c:/r/who/Morticd10_part1")
file2 = fread("c:/r/who/Morticd10_part2")
all_data = rbind(file1,file2)
focus_data = all_data[Country %in% c(1430) & Cause != "AAA"]
# coding as per CDC
coding = fread("c:/r/motor.csv", header=T)
coding = rbind(coding, data.frame(Code = c("V89","V09", "V19")))
focus_data[Cause %in% coding$Code, Type := "Traffic"]
focus_data[!Cause %in% coding$Code, Type := "Non-Traffic"]
focus_data[,Major:=str_sub(Cause,1,1)]
focus_data[Major == "V"][,sum(Deaths1), keyby = .(Type, Cause)]
focus_data = focus_data[,c(1,4,7,10:35,40),with=F]
focus_data = focus_data %>% melt(id.vars=c("Country", "Year", "Sex", "Type")) %>% data.table()
# build a look-up table to deal with formatting issues
variable_names = focus_data[,.(variable = unique(variable))]
variable_names[,group := as.integer(str_replace(variable, "Deaths", ""))]
variable_names=variable_names[group != 1]
variable_names=variable_names[group != 26]
variable_names[,group:=group-2]
variable_names[group ==0,age_lower:=0]
variable_names[group <=4 & group >0,age_lower:=1]
variable_names[group >4,age_lower:=5+(group-5)*5]
variable_names[group ==0,age_group:=paste0(0, "-", 1)]
variable_names[group <=4 & group >0,age_group:=paste0(1, "-", 4)]
variable_names[group >4,age_group:=paste0(age_lower, "-", age_lower+4)]
### Derive the number of deaths due to motor accident reported in the WHO data
focus_data %>% setkey(variable)
variable_names %>% setkey(variable)
focus_data = focus_data %>% merge(variable_names)
death_data_cause = focus_data[,.(Deaths = sum(value)), keyby = .(Country, Year, Sex, Type, age_lower, age_group )]
death_data_total = death_data_cause[,.(Deaths = sum(Deaths)), keyby = .(Country, Year, Sex, age_lower, age_group )]
death_data_total[, Type:="Total"]
death_data = rbind(death_data_cause, death_data_total)
death_data = death_data[Sex !=9]
death_data[Type != "Total"] %>% ggplot(aes(x=age_lower, y = Deaths, colour = Year)) + geom_point(aes(colour = Year))+ facet_wrap(Sex~Type, scales = "free_y")+
geom_line(aes(group = Year))
death_data_factors = death_data %>% dcast.data.table(Country + Year + Sex + age_group+ age_lower~Type, value.var = "Deaths")
death_data_factors[,reduction := 1-`Non-Traffic`/Total]
death_data_factors[,Country_Name := "South Africa"]
death_data_factors %>% ggplot(aes(x=age_lower, y = reduction, colour = Year)) + geom_point(aes(colour = Year))+ facet_grid(Sex~Country_Name)+
geom_line(aes(group = Year))
ggsave("c:/r/reduction_pre_corr.jpg")
### check age pattern of deaths
age_pattern = death_data_cause[Type == "Traffic"] %>% setkey(Country, Year, Sex, Type, age_lower, age_group)
age_pattern[,total_deaths:=sum(Deaths),keyby = .(Country, Year, Sex)]
age_pattern[,propn_deaths:=Deaths/total_deaths]
age_pattern = age_pattern[,c(2,3,5,9),with=F]
age_pattern %>% setkey(Year, Sex, age_lower)
age_pattern[Sex != 9] %>% ggplot(aes(x=age_lower, y=propn_deaths, colour = Year)) + facet_grid(Sex~.)+
geom_line(aes(group = Year))
ggsave("c:/r/agepattern.jpg")
focus_data = all_data[Country %in% c(1430, 2450,4308) & Cause != "AAA"]
focus_data = focus_data[,c(1,4,7,10:35,6),with=F]
focus_data = focus_data %>% melt(id.vars=c("Country", "Year", "Sex", "Cause")) %>% data.table()
focus_data[Cause %in% coding$Code, Type := "Traffic"]
focus_data[!Cause %in% coding$Code, Type := "Non-Traffic"]
focus_data = focus_data[Type != "Traffic"]
focus_data %>% setkey(variable)
variable_names %>% setkey(variable)
focus_data = focus_data %>% merge(variable_names)
focus_data = focus_data[!is.na(value),.(Deaths = sum(value)), keyby = .(Country, Year, Sex, Cause, age_lower, age_group )]
focus_data[,total_deaths_Cause:=sum(Deaths),keyby = .(Country, Year, Sex, Cause)]
focus_data[,propn_deaths_Cause:=Deaths/total_deaths_Cause]
focus_data = focus_data[Sex !=9]
focus_data %>% setkey(Year, Sex, age_lower)
focus_data = merge(focus_data, age_pattern)
top20 = focus_data[,.(Deaths = sum(Deaths), distance = sqrt(sum(abs(propn_deaths_Cause-propn_deaths)^2))), keyby = .(Country, Cause)][!is.na(distance)& Deaths>10][order(distance)][1:20]
top20%>% fwrite("c:/r/closest_match.csv")
focus_data[Country == 1430 & Cause == "Y34"] %>% ggplot(aes(x=age_lower, y = propn_deaths_Cause, colour = Year)) + facet_grid(Sex~.)+
geom_line(aes(group = Year)) +
geom_point(data = focus_data[Country == 1430 & Cause == "Y34"],aes(x=age_lower, y = propn_deaths, colour = Year))
ggsave("c:/r/agepattern_match.jpg")
### Compare totals to other sources
summary = death_data[!is.na(Deaths) & Type == "Traffic",.(Deaths = sum(Deaths)), keyby = .(Year)]
rtmc = fread("c:/users/user/dropbox/traffic_deaths/rtmc stats.csv") %>% setkey(Year)
rtmc = merge(summary,rtmc)
rtmc[,propn:=Deaths/`Crash Fatalities`]
rtmc %>% fwrite("c:/r/rtmc.csv")
### Correct the WHO deaths using the RTMC reported deaths
death_data_factors = death_data[!is.na(Deaths)] %>% dcast.data.table(Country + Year + Sex + age_group+ age_lower~Type, value.var = "Deaths")
rtmc %>% setkey(Year)
death_data_factors %>% setkey(Year)
death_data_factors = death_data_factors %>% merge(rtmc)
death_data_factors[,extra := Traffic/propn - Traffic]
death_data_factors[,Traffic := Traffic + extra]
death_data_factors[,`Non-Traffic` := `Non-Traffic` - extra]
death_data_factors[,reduction := 1-`Non-Traffic`/Total]
death_data_factors[,Country_Name := "South Africa"]
death_data_factors %>% ggplot(aes(x=age_lower, y = reduction, colour = Year)) + geom_point(aes(colour = Year))+ facet_grid(Sex~Country_Name)+
geom_line(aes(group = Year))
ggsave("c:/r/sa_reduction.jpg")
### derive mortality rates
thembisa = fread("c:/r/thembisa.csv", header=T) %>% melt(id.vars = c("Sex", "Age"))
thembisa[,Year:=as.integer(as.character(variable))]
thembisa[,qx:=value]
thembisa[,variable:=NULL]
thembisa[,value:=NULL]
thembisa[,mx:=-log(1-qx)]
thembisa[,log_mx:=log(mx)]
extend_gompertz = function(dat){
dat = dat %>% data.table()
dat = dat[Age >= 80]
fit = lm(log_mx~Age,data=dat)
pred_dat = expand.grid(Age = seq(90,110), Year =dat$Year[1], Sex = dat$Sex[1])%>% data.table()
pred_dat$log_mx = predict(fit, pred_dat)
pred_dat[,mx := exp(log_mx)]
pred_dat[,qx := 1- exp(-mx)]
return(pred_dat)
}
to_110 = dlply(thembisa, .(Year,Sex), extend_gompertz) %>% rbindlist()
thembisa = rbind(thembisa[Age<91], to_110) %>% setkey(Year, Sex, Age)
thembisa[Age==0, age_group := paste0(0, "-", 1)]
thembisa[Age < 5 & Age>0, age_group := paste0(1, "-", 4)]
thembisa[Age >= 5,age_group:=paste0(trunc(Age/5)*5, "-", trunc(Age/5)*5+4)]
thembisa[Age >= 100,age_group:=paste0(95, "-", 99)]
death_data_factors = death_data_factors[,c(1,3,4,13),with=F] %>% setkey(Year, Sex, age_group)
thembisa %>% setkey( Year, Sex, age_group)
mortality_rates = merge(thembisa,death_data_factors)
mortality_rates[,qx_no_traffic:= (1-reduction)*qx]
mortality_rates[,px:=1-qx]
mortality_rates[,px_no_traffic:=1-qx_no_traffic]
mortality_rates[,tpx := c(1,cumprod(px)), by = .(Year, Sex)]
mortality_rates[,tpx_no_traffic := c(1,cumprod(px_no_traffic)), by = .(Year, Sex)]
mortality_rates[,Lx:=(tpx+lead(tpx))/2,by = .(Year, Sex)]
mortality_rates[,Lx:=ifelse(is.na(Lx),0,Lx),by = .(Year, Sex)]
mortality_rates[,Tx:=(rev(cumsum(rev(Lx)))),by = .(Year, Sex)]
mortality_rates[,ex:=Tx/Lx]
mortality_rates[,Lx_no_traffic:=(tpx_no_traffic+lead(tpx_no_traffic))/2,by = .(Year, Sex)]
mortality_rates[,Lx_no_traffic:=ifelse(is.na(Lx_no_traffic),0,Lx_no_traffic),by = .(Year, Sex)]
mortality_rates[,Tx_no_traffic:=(rev(cumsum(rev(Lx_no_traffic)))),by = .(Year, Sex)]
mortality_rates[,ex_no_traffic:=Tx_no_traffic/Lx_no_traffic]
mortality_rates[,Country_Name := "South Africa"]
mortality_rates %>% ggplot(aes(x=Age, y=log(qx))) + geom_line(aes(colour = Year, group = Year))+ facet_wrap(Sex~Country_Name)
mortality_rates[Year==2015] %>% ggplot(aes(x=Age, y=log(qx))) +
geom_line(aes(group = Year))+
facet_grid(Country_Name +Sex~Year)+
geom_line(aes(colour = "No Motor", group = Year, x=Age, y= log(qx_no_traffic)), data =mortality_rates[Year==2015] )
ggsave("c:/r/qx.jpg")
e0 = mortality_rates[Age == 0][,c("Country_Name", "Sex", "Year", "ex", "ex_no_traffic"),with=F][,Increase:=ex_no_traffic-ex] %>% setkey(Country_Name, Sex, Year)
e0 %>% write.table("c:/r/eo_sa.csv", sep=",", row.names=F)
# Years of life saved if no traffic accidents. Note that births registered at 90% completeness
475355 * e0[Sex == 1 & Year==2015]$Increase/.9
468447 * e0[Sex == 2 & Year==2015]$Increase/.9