dataFolder <- "./data/UCI HAR Dataset/"

activityLabels <- read.table(paste0(dataFolder, "activity_labels.txt"), sep=" ",

stringsAsFactors=FALSE)

colnames(activityLabels) <- c("ActivityID", "ActivityName")

activityLabels$ActivityName <- gsub("_", " ", activityLabels$ActivityName)

data <- list()

for ( dir in c("test", "train") ) {

folder <- paste0(dataFolder, dir)

data$subject <- rbind(data$subject, read.table(paste0(folder, "/subject_",

dir, ".txt")))

data$activity <- rbind(data$activity, read.table(paste0(folder, "/y_", dir,

".txt")))

data$X <- rbind(data$X, read.table(paste0(folder, "/X_", dir, ".txt")))

}

data$feature <- read.table(paste0(dataFolder, "features.txt"),

stringsAsFactors=FALSE)

names(data$X) <- data$feature[, 2]

names(data$subject) <- "SubjectID"

names(data$feature) <- c("FeatureID", "Feature")

activityNames <- c() # initializing

for ( i in 1:nrow(data$activity))

activityNames <- c(activityNames, activityLabels[data$activity[i, ], 2])

data$activity <- data.frame(data$activity, activityNames, stringsAsFactors=F)

names(data$activity) <- c("ActivityID", "ActivityName")

names(data)

str(data)

library(plyr)

features <- data$feature$Feature

featuresToGo <- grep("^t.+mean\\(\\)|^t.+std\\(\\)", features)

data <- data.frame(data$subject, data$activity$ActivityName,

data$X[, featuresToGo], stringsAsFactors=FALSE)

names(data)[c(1,2)] <- c("subject", "activity")

features <- features[featuresToGo]

features <- gsub("\\(|\\)|-", "", features)

features <- gsub("mean", "Mean", features)

features <- gsub("std", "Std", features)

names(data)[3:ncol(data)] <- features

averages <- ddply(data, .(subject, activity), numcolwise(mean))

write.csv(averages, "averages.txt", row.names=FALSE)