adding more standard deviations to a plot - r

Question

I'm trying to add more standard deviation to my current plot. I need to add 1std and 3std, I've already addeed the 2std to my plot.
This is my code:
tidyverse_downloads_rollmean <- treasury %>%
tq_mutate(
# tq_mutate args
select = yield,
mutate_fun = rollapply,
# rollapply args
width = 360,
align = "right",
FUN = mean,
# mean args
na.rm = TRUE,
# tq_mutate args
col_rename = "mean_360"
)
This is for the 2std but I need to add 1std and 3std in the same plot:
custom_stat_fun_2 <- function(x, na.rm = TRUE) {
m <- mean(x, na.rm = na.rm)
s <- sd(x, na.rm = na.rm)
hi <- m + 2*s
lo <- m - 2*s
ret <- c(mean = m, stdev = s, hi.95 = hi, lo.95 = lo)
return(ret)
}
I added to my data:
rollstats<- treasury %>%
tq_mutate(
select = yield,
mutate_fun = rollapply,
# rollapply args
width = 360,
align = "right",
by.column = FALSE,
FUN = custom_stat_fun_2,
# FUN args
na.rm = TRUE
)
This is my plot:
rollstats %>%
ggplot(aes(x = date)) +
# Data
geom_line(aes(y = yield), color = "grey40", alpha = 0.5, size =1) +
geom_ribbon(aes(ymin = lo.95, ymax = hi.95), alpha = 0.4) +
geom_point(aes(y = mean), linetype = 2, size = 0.5, alpha = 0.5) +
# Aesthetics
labs(title = "tidyverse packages: Volatility and Trend", x = "",
subtitle = "360-Day Moving Average with 95% Confidence Interval Bands (+/-2 Standard Deviations)") +
scale_color_tq(theme = "light") +
theme_tq() +
theme(legend.position="none")
This is my output:
But I want something like this:
So how can I add the 1std and 2std? Is there another way to plot 1std, 2std and 3std in the same plot? Thanks in advance!

Answer 1

You haven't provided a reprex so hard to help you. Tidyquant has functions to plot the standard deviation bands for you (geom_bbands). But, here's an idea with only ggplot2 using different data. Calculate the 1st, 2nd, and 3rd standard deviations:
library(tidyquant)
custom_stat_fun_2 <- function(x, na.rm = TRUE) {
m <- mean(x, na.rm = na.rm)
s <- sd(x, na.rm = na.rm)
hi1 <- m + s
lo1 <- m - s
hi2 <- m + 2*s
lo2 <- m - 2*s
hi3 <- m + 3*s
lo3 <- m - 3*s
ret <- c(mean = m, stdev = s,hi1 = hi1, lo1 = lo1, hi2=hi2, lo2=lo2, hi3=hi3, lo3=lo3)
return(ret)
}
treasury <- treasuryTR::get_yields("DGS10", format_out = "tibble")
rollstats<- treasury |>
tq_mutate(
select = DGS10,
mutate_fun = rollapply,
# rollapply args
width = 360,
align = "right",
by.column = FALSE,
FUN = custom_stat_fun_2
) |>
na.omit()
Melt the data frame to have one column for hi and one for lo and then set factor levels so they plot in reverse order:
rollsds <- tidyr::pivot_longer(rollstats,cols = starts_with(c("hi", "lo")),
names_to = c(".value", "sd"), names_pattern = "(.*)(\\d)")
rollsds$sd <- factor(as.character(rollsds$sd), levels=c(3,2,1))
Plot
library(ggplot2)
rollstats |>
ggplot(aes(x = date)) +
# Data
geom_ribbon(data=rollsds, aes(ymax = hi, ymin=lo, fill=sd, color=sd), alpha=0.3) +
geom_line(aes(y = mean), linetype = 2, size = 0.5, alpha = 0.5) +
geom_line(aes(y = DGS10), color = "midnightblue", alpha = 0.7, size =1) +
# Aesthetics
theme_tq()