SocialDataBlog’s kind reference in post Horizon plots with ggplot (not) motivated me to finish what the post started. I knew that ggplot2 would be a little more difficult to use for the purpose of a horizon plot, but I felt compelled to provide at least one example of a horizon plot for each of the major R graphing packages. I achieved a good result but the code is not as elegant or as flexible as I would like. Readers more comfortable with ggplot2, please bash, fork, and improve.
 |
| From TimelyPortfolio |
R code in GIST (do raw for copy/paste):
#first attempt at implementing horizon plots in ggplot2
#pleased with result but code sloppy and inflexible
#as always very open to improvements and forks
require(ggplot2)
require(reshape2)
require(quantmod)
require(PerformanceAnalytics)
require(xtsExtra)
data(edhec)
origin = 0
horizonscale = 0.1
#get 12 month rolling return of edhec indexes
roc <- as.xts(apply(cumprod(edhec+1),MARGIN=2,ROC,n=12,type="discrete"),order.by=index(edhec))
roc.df <- as.data.frame(cbind(index(roc),coredata(roc)))
roc.melt <- melt(roc.df,id.vars=1)
roc.melt[,1] <- as.Date(roc.melt[,1]) #convert back to a Date
horizon.panel.ggplot <- function(df, title) {
#df parameter should be in form of date (x), grouping, and a value (y)
colnames(df) <- c("date","grouping","y")
#get some decent colors from RColorBrewer
#we will use colors on the edges so 2:4 for red and 7:9 for blue
require(RColorBrewer)
col.brew <- brewer.pal(name="RdBu",n=10)
#get number of bands for the loop
#limit to 3 so it will be much more manageable
nbands = 3
#loop through nbands to add a column for each of the positive and negative bands
for (i in 1:nbands) {
#do positive
df[,paste("ypos",i,sep="")] <- ifelse(df$y > origin,
ifelse(abs(df$y) > horizonscale * i,
horizonscale,
ifelse(abs(df$y) - (horizonscale * (i - 1) - origin) > origin, abs(df$y) - (horizonscale * (i - 1) - origin), origin)),
origin)
#do negative
df[,paste("yneg",i,sep="")] <- ifelse(df$y < origin,
ifelse(abs(df$y) > horizonscale * i,
horizonscale,
ifelse(abs(df$y) - (horizonscale * (i - 1) - origin) > origin, abs(df$y) - (horizonscale * (i - 1) - origin), origin)),
origin)
}
#melt data frame now that we have added a column for each band
#this will fit ggplot2 expectations and make it much easier
df.melt <- melt(df[,c(1:2,4:9)],id.vars=1:2)
#name the columns for reference
#try to be generic
colnames(df.melt) <- c("date","grouping","band","value")
#use ggplot to produce an area plot
p <- ggplot(data=df.melt) +
geom_area(aes(x = date, y = value, fill=band),
#alpha=0.25,
position="identity") + #this means not stacked
scale_fill_manual(values=c("ypos1"=col.brew[7], #assign the colors to each of the bands; colors get darker as values increase
"ypos2"=col.brew[8],
"ypos3"=col.brew[9],
"yneg1"=col.brew[4],
"yneg2"=col.brew[3],
"yneg3"=col.brew[2])) +
ylim(origin,horizonscale) + #limit plot to origin and horizonscale
facet_grid(grouping ~ .) + #do new subplot for each group
theme_bw() + #this is optional, but I prefer to default
opts(legend.position = "none", #remove legend
strip.text.y = theme_text(),#rotate strip text to horizontal
axis.text.y = theme_blank(),#remove y axis labels
axis.ticks = theme_blank(), #remove tick marks
axis.title.y = theme_blank(),#remove title for the y axis
axis.title.x = theme_blank(),#remove title for the x axis
title = title, #add a title from function parameter
plot.title = theme_text(size=16, face="bold", hjust=0)) #format title
return(p)
}
horizon.panel.ggplot(roc.melt, "EDHEC Indexes Return (Rolling 1 Year)")