Correlation Sample

corr.r

require(chron)
require(Hmisc)
require(ggplot2)

#this function creates and returns date time for indexing purposes later
f = function(d, t) as.chron(paste(strptime(d, "%Y.%m.%d"), t))
#path to data, change for pointing your repository
path <- "G://TickData//" 
#suffix of the file names with extension
suffix <- "_UTC+0_00_noweekends_fxmtf.csv" 
#timeframe
tf <- "H1"
#in following lines data are read into the memory, note we are loading just Close price
#this could be done more effectivelly,
#but I'd like to leave soemthing for some upcoming articles :)
AUDUSD <- read.zoo(read.csv(file=paste(path, "AUDUSD_", tf, suffix, sep = ""), header=F, col.names=c("Date", "Time", "O","H","L","C","V")), header=F, index=1:2, FUN=f)$C
EURUSD <- read.zoo(read.csv(file=paste(path, "EURUSD_", tf, suffix, sep = ""), header=F, col.names=c("Date", "Time", "O","H","L","C","V")), header=F, index=1:2, FUN=f)$C
GBPUSD <- read.zoo(read.csv(file=paste(path, "GBPUSD_", tf, suffix, sep = ""), header=F, col.names=c("Date", "Time", "O","H","L","C","V")), header=F, index=1:2, FUN=f)$C
NZDUSD <- read.zoo(read.csv(file=paste(path, "NZDUSD_", tf, suffix, sep = ""), header=F, col.names=c("Date", "Time", "O","H","L","C","V")), header=F, index=1:2, FUN=f)$C
USDCAD <- read.zoo(read.csv(file=paste(path, "USDCAD_", tf, suffix, sep = ""), header=F, col.names=c("Date", "Time", "O","H","L","C","V")), header=F, index=1:2, FUN=f)$C
USDJPY <- read.zoo(read.csv(file=paste(path, "USDJPY_", tf, suffix, sep = ""), header=F, col.names=c("Date", "Time", "O","H","L","C","V")), header=F, index=1:2, FUN=f)$C
USDCHF <- read.zoo(read.csv(file=paste(path, "USDCHF_", tf, suffix, sep = ""), header=F, col.names=c("Date", "Time", "O","H","L","C","V")), header=F, index=1:2, FUN=f)$C
#merge everything into the single data frame
df <- merge(EURUSD, AUDUSD, GBPUSD, NZDUSD, USDCAD, USDJPY, USDCHF)
#calculate correlation, tail with parameter 11 instructs runtimes to get only last 11 observations
cor <- rcorr(tail(df,11), type="pearson")
#ggplot likes the data 'melted' one value per row
m <-melt(cor$r) 
#plot it finally
ggplot(data=m, aes(x=Var1, y=Var2, fill=value)) + geom_tile()