run_analysis.R

## Coursera Getting and Cleaning Data Course Project
## Jianglin Huang
## 2016-03-15

rm(list=ls())
# set the working destination
#setwd("C:/Users/jianhuang7/Dropbox/WorkStation/PhD/Course_Materials/Coursera/R/UCI HAR Dataset")
setwd("/Users/Hellwalker_Mac/Dropbox/WorkStation/PhD/Course_Materials/Coursera/R/UCI HAR Dataset")

#imports features.txt
fs = read.table('./features.txt',header=F) 
#imports activity_labels.txt
activityType = read.table('./activity_labels.txt',header=FALSE) 
subjectTrain = read.table('./train/subject_train.txt',header=FALSE)#imports subject_train.txt
xTrain       = read.table('./train/x_train.txt',header=FALSE)#imports x_train.txt
yTrain       = read.table('./train/y_train.txt',header=FALSE) #imports y_train.txt

# 1Merges the training and the test sets to create one data set.
# Assigin column names to the data imported above
colnames(activityType)  = c('activityId','activityType');
colnames(subjectTrain)  = "subjectId";
colnames(xTrain)        = fs[,2]; 
colnames(yTrain)        = "activityId";

# cCreate the final training set by merging yTrain, subjectTrain, and xTrain
trainingData = cbind(yTrain,subjectTrain,xTrain);

# Read in the test data
subjectTest = read.table('./test/subject_test.txt',header=FALSE); #imports subject_test.txt
xTest       = read.table('./test/x_test.txt',header=FALSE); #imports x_test.txt
yTest       = read.table('./test/y_test.txt',header=FALSE); #imports y_test.txt

# Assign column names to the test data imported above
colnames(subjectTest) = "subjectId";
colnames(xTest)       = fs[,2]; 
colnames(yTest)       = "activityId";

# Create the final test set by merging the xTest, yTest and subjectTest data
testData = cbind(yTest,subjectTest,xTest);

# Combine training and test data to create a final data set
finalData = rbind(trainingData,testData);
# to select the desired mean() & stddev() columns
colNames  = colnames(finalData); 

# 2. Extract only the measurements on the mean and standard deviation for each measurement. 
logicalVector = (grepl("activity..",colNames) | grepl("subject..",colNames) | grepl("-mean..",colNames) & !grepl("-meanFreq..",colNames) & !grepl("mean..-",colNames) | grepl("-std..",colNames) & !grepl("-std()..-",colNames));

# Subset finalData table based on the logicalVector to keep only desired columns
finalData = finalData[logicalVector==TRUE];

# 3. Use descriptive activity names to name the activities in the data set
# Merge the finalData set with the acitivityType table to include descriptive activity names
finalData = merge(finalData,activityType,by='activityId',all.x=TRUE);

# Updating the colNames vector to include the new column names after merge
colNames  = colnames(finalData); 

# 4. Appropriately label the data set with descriptive activity names. 
for (i in 1:length(colNames)) 
{
  colNames[i] = gsub("\\()","",colNames[i])
  colNames[i] = gsub("-std$","StdDev",colNames[i])
  colNames[i] = gsub("-mean","Mean",colNames[i])
  colNames[i] = gsub("^(t)","time",colNames[i])
  colNames[i] = gsub("^(f)","freq",colNames[i])
  colNames[i] = gsub("([Gg]ravity)","Gravity",colNames[i])
  colNames[i] = gsub("([Bb]ody[Bb]ody|[Bb]ody)","Body",colNames[i])
  colNames[i] = gsub("[Gg]yro","Gyro",colNames[i])
  colNames[i] = gsub("AccMag","AccMagnitude",colNames[i])
  colNames[i] = gsub("([Bb]odyaccjerkmag)","BodyAccJerkMagnitude",colNames[i])
  colNames[i] = gsub("JerkMag","JerkMagnitude",colNames[i])
  colNames[i] = gsub("GyroMag","GyroMagnitude",colNames[i])
};

# Reassigning the new descriptive column names to the finalData set
colnames(finalData) = colNames;

# 5. Create a second, independent tidy data set with the average of each variable for each activity and each subject. 
finalDataNoActivityType  = finalData[,names(finalData) != 'activityType'];

# Summarizing the finalDataNoActivityType table to include just the mean of each variable for each activity and each subject
tidyData    = aggregate(finalDataNoActivityType[,names(finalDataNoActivityType) != c('activityId','subjectId')],by=list(activityId=finalDataNoActivityType$activityId,subjectId = finalDataNoActivityType$subjectId),mean);
tidyData    = merge(tidyData,activityType,by='activityId',all.x=TRUE);

write.table(tidyData, './tidyData.txt',row.names = F,sep='\t')