run_analysis.R

# Course Project
# Coursera Getting and Cleaning Data
# Johns Hopkins University
# Author: Ben McKibben
# 2015-04-24
# https://github.com/bmckibben/Tidy_Data_Project/blob/master/run_analysis.R
# Requirements
#
# 1. Merges the training and the test sets to create one data set.
# 2. Extracts only the measurements on the mean and standard deviation for each measurement.
# 3. Uses descriptive activity names to name the activities in the data set
# 4. Appropriately labels the data set with descriptive activity names.
# 5. Creates a second, independent tidy data set with the average of each variable for each activity and each subject.
#=============================================================================
# STEP ZERO: Prepare environment
#=============================================================================
if (!require("data.table")) {
  install.packages("data.table")
}

if (!require("reshape2")) {
  install.packages("reshape2")
}

if (!require("plyr")) {
  install.packages("plyr")
}

require("data.table")
require("reshape2")
require("plyr")
#=============================================================================
# STEP ONE: Merges the training and the test sets to create one data set.
#=============================================================================
# Read Test Data Files
 testSubject <- read.table('./UCI HAR Dataset/test/subject_test.txt')
 testX <- read.table('./UCI HAR Dataset/test/x_test.txt')
 testY <- read.table('./UCI HAR Dataset/test/y_test.txt')

# create a set column for DF
 set <- 'test'
 #merge first two dt's and identify set. using cbind so rows remain in original sort
 testSY <- cbind(testSubject,set,testY)
 #merge third dt's
 test <- cbind(testSY,testX)

# Read Train Data Files
 trainSubject <- read.table('./UCI HAR Dataset/train/subject_train.txt')
 trainX <- read.table('./UCI HAR Dataset/train/x_train.txt')
 trainY <- read.table('./UCI HAR Dataset/train/y_train.txt')

 # create a set column for DF
 set <- 'train'
 #merge first two dt's and identify set. using cbind so rows remain in original sort
 trainSY <- cbind(trainSubject,set,trainY)
 #merge third dt's
 train<- cbind(trainSY,trainX)

#merge the test and train data tables
 c3pd <- rbind(test,train)
# clean up the environment
rm(test,testSubject,testY,testX,testSY,train,trainSubject,trainY,trainX,trainSY)
#=============================================================================
# STEP TWO: Extracts only the measurements on the mean and standard deviation 
#           for each measurement.
#=============================================================================

#read feature column  names
 colNames <- read.table('./UCI HAR Dataset/features.txt',stringsAsFactors=FALSE)[[2]]

  colNames <- c("subject","set","activity", colNames)

 #set remaining column names from features.txt
  names(c3pd) <- colNames

# discard data columns without "mean' or "std"
colNames <- names(c3pd)
selectedCols <- c("subject" ,"set", "activity", grep("-(mean|std)\\(\\)", colNames, value=TRUE))
c3pd <- c3pd[,selectedCols]

#=============================================================================
# STEP THREE: Uses descriptive activity names to name the activities in the data set
#=============================================================================
# replace value in the activity column with their text equivelents from activity_labels.txt
c3pd$activity[c3pd$activity==1] <- "walking"
c3pd$activity[c3pd$activity==2] <- "walking upstairs"
c3pd$activity[c3pd$activity==3] <- "walking downstairs"
c3pd$activity[c3pd$activity==4] <- "sitting"
c3pd$activity[c3pd$activity==5] <- "standing"
c3pd$activity[c3pd$activity==6] <- "laying"

#=============================================================================
# STEP FOUR Appropriately labels the data set with descriptive activity names.
#=============================================================================

# translate column names
colNames <- names(c3pd)
colNames <- gsub(pattern="^t",replacement="time",x=colNames)
colNames <- gsub(pattern="^f",replacement="freq",x=colNames)
colNames <- gsub(pattern="-?mean[(][)]-?",replacement="Mean",x=colNames)
colNames <- gsub(pattern="-?std[()][)]-?",replacement="Std",x=colNames)
colNames <- gsub(pattern="-?meanFreq[()][)]-?",replacement="MeanFreq",x=colNames)
colNames <- gsub(pattern="BodyBody",replacement="Body",x=colNames) 
names(c3pd) <- colNames

#=============================================================================
# STEP FIVE: Creates a second, independent tidy data set with the average of 
#            each variable for each activity and each subject.
#=============================================================================

# calc average mean and export tidy data set

id_labels   = c("subject","set","activity")
data_labels = setdiff(colnames(c3pd), id_labels)
melt_data      = melt(c3pd, id = id_labels, measure.vars = data_labels)

tidy_data   = dcast(melt_data, subject + activity ~ variable, mean)
write.table(tidy_data, file = "./tidy_data.txt")