Taking the data below, and going the long way round, I make a plot of probabilities with confidence intervals.
I am looking next two create a red vertical line between the CI_U and CI_L points that have been plotted for each X.
Other than manually giving a abline command per x, can this be automated
Here is some data to follow:
set.seed(123)
x <- sample(1:10, 2000, replace = T)
y <- c(rep(0.1, times = 500),rep(0.25, times = 500),rep(0.4, times = 500),rep(0.85, times = 500))
z <- rbinom(n=2000, size=1, prob=y)
df1 <- data.frame(x,z)
df1$x <- as.factor(df1$x)
# ---------------------------------------------------------------------------------------------------------------------------------
# 1. we want the counts of success per interval
success<-(1:10)
for (i in 1:10) {
success[i] <- (sum(df1$z[df1$x==success[i]]))
}
success
# 1. we want the counts of population in each two year interval
population<-(1:10)
for (i in 1:10) {
population[i] <- length(df1$z[df1$x==population[i]])
}
population
X <- c(1:10)
df2 <- data.frame(X, success, population)
df2$pred <- df2$success/df2$population
# in a similiar manner, calculate the ratio of those injured to not, per vehicle year
CI_L<-(1:10)
for (i in 1:10) {
CI_L[i] <- binom.test(df2$success[i], df2$population[i], df2$pred[i])[[4]][[1]]
}
# in a similiar manner, calculate the ratio of those injured to not, per vehicle year
CI_U<-(1:10)
for (i in 1:10) {
CI_U[i] <- binom.test(df2$success[i], df2$population[i], df2$pred[i])[[4]][[2]]
}
df2$CI_U <- CI_U
df2$CI_L <- CI_L
# create our plot
plot(df2$X, df2$pred, ylim = c(0,1))
points(df2$X, df2$CI_L, pch='-',, add = T, col='red')
points(df2$X, df2$CI_U, pch='-', add = T, col='red')
