I’ve spent years honing my economic acumen and valuation modeling techniques, but what good are
those skills if I don’t have an easy and efficient way to access the large volume of data needed to
run meaningful analysis? I’m going to teach you a method to overcome one of he biggest hurdles to
getting started: Actually getting the data to conduct research. I developed a system to automate this
process by creating a function that utilizing the tseries package in R to cache historical security
prices onto my local drive, which effectively “stages” the data in order to run the actual analysis.
Much of the functionality here was developed by using Financial Analytics with R by Bennett & Hugen as a guide.
The Following code will create a sub-directory in your source file location on your local drive
All you need to do is feed the function a vector of ticker symbols (datatype must be as.character). You can specify the start/end dates, but it defaults to the trailing 12 months. Downloading the 5 year historical returns of all 505 securities in the S&P 500 took sub 3.5 minutes on my machine.
library(tseries)
create_price_dir <- function(labels, start_date=Sys.Date()-365, end_date=Sys.Date()-1) {
path <- dirname(rstudioapi::getSourceEditorContext()$path)
setwd(path)
#days <- length(get.hist.quote('AAPL', start = start_date, end = end_date, quote = 'Adj'))
len <- length(labels)
try(dir.create("data"))
k <- 1
for (i in labels) {
try({
prices <- as.vector(get.hist.quote(i, start = start_date, end = end_date, quote = 'Adj'))
#print(prices)
dates <- time(get.hist.quote(i, start = start_date, end = end_date, quote = 'Adj'))
bound <- cbind(as.character(dates), prices)
file_name <- paste(path, "/data/", "cached", i, ".csv", sep = '')
write.csv(bound, file = file_name, row.names = F)
print(paste(k, "of", len, "done", sep = " "))
k <- k+1
}, silent = T)
}
return(paste("Done with", len-k, "errors"))
}
