Create one R script run_analysis.R to do the following:

1. Merge the training and the test sets to create one data set.
2. Extract only the measurements on the mean and standard deviation for each measurement.
3. Use descriptive activity names to name the activities in the data set
4. Appropriately labels the data set with descriptive feature names.
5. Create a second, independent tidy data set with the average of each variable for each activity and each subject.

• The required dataset was downloaded and unzipped in a local directory using the following R script

  get_zip_file <- function(zipdir, url) 
  {            
    #Create a temp file to save the downloaded zipped file          
    temp <- tempfile()   
    
    #download the file from the given url         
    download.file(url, temp) 
    
    #Create the zipdir directory         
    dir.create(zipdir)    
    
    #Unzip the file into the directory         
    unzip(temp, exdir = zipdir)     
  }
  

• Local directory is assigned to zipdir in order to access all the relevant files.

  zipdir<-"UCI HAR Dataset"
  

• Observed Data from train/X_train.txt and test/X_test.txt is read into data frames data_Train and data_Test respectively from appropriate folders in zipdir,

  data_Train<-read.table(paste(zipdir,"train/X_train.txt", sep= "/"), stringsAsFactors = FALSE)
  data_Test<-read.table(paste(zipdir,"test/X_test.txt", sep= "/"), stringsAsFactors = FALSE)
  

• Subject Info from train/subject_train.txt and test/subject_test.txt is read into data frames data_SubTrain and data_SubTest respectively,

  data_SubTrain<-read.table(paste(zipdir,"train/subject_train.txt", sep= "/"), stringsAsFactors = FALSE)
  data_SubTest<-read.table(paste(zipdir,"test/subject_test.txt", sep= "/"), stringsAsFactors = FALSE)       
  

• Labels from train/y_train.txt and test/y_test.txt is read into data frames data_Train_Activity and data_Test_Activity respectively,

  data_Train_Activity<-read.table(paste(zipdir,"train/y_train.txt", sep= "/"), stringsAsFactors = FALSE)
  data_Test_Activity<-read.table(paste(zipdir,"test/y_test.txt", sep= "/"), stringsAsFactors = FALSE)
  

• List of all features from features.txt is read into data frame data_features. These are the 561 observed variables in the Train and Test data sets.

  data_features<-read.table(paste(zipdir,"features.txt", sep= "/"), stringsAsFactors = FALSE)
  

• Activity name for each activity, which has been provided in the file activity_labels.txt, is loaded into to data frame data_activity_labels, which will be used in Step 3.

  data_activity_labels<-read.table(paste(zipdir,"activity_labels.txt", sep= "/"), stringsAsFactors = FALSE)        
  

• Data frame data_Train and data_Test are merged together using rbind() into data frame data_Observations.

  data_Observations<-rbind(data_Train, data_Test)
  

• Data frame data_SubTrain and data_SubTest are merged together using rbind() into data frame data_Subject.

  data_Subject<-rbind(data_SubTrain,data_SubTest)
  

• Data frame data_Train_Activity and data_Test_Activity are merged together using rbind() into data frame data_Activity.

  data_Activity<-rbind(data_Train_Activity, data_Test_Activity)
  

• Column names for data frame data_Observations is assigned as

  setnames(data_Observations,names(data_Observations), data_features[,2])
  

• Column names for data_SubTrain and data_Train_Activity is also assigned as

  setnames(data_Subject, names(data_Subject), "subject")         
  setnames(data_Activity, names(data_Activity), "activity")
  

• As only the measurements on the mean and standard deviation for each measurement is required, grepl() is used to extract these relevant columns.

• grepl() is used get only those columns with mean() and std() at the end. This gives a logical vector of column names selectColumns only with mean() and std() as required.

  selectColumns<-(grepl("-mean\\()$",names(data_Observations)) &
              !grepl("-meanFreq\\()",names(data_Observations)) | 
               grepl("-std\\()$",names(data_Observations)))
  

• selectColumns is used to extract only the relevant columns from the data_Observations data frame.

  data_Observations<-data_Observations[,selectColumns]
  

• Now, all three data frames are ready to be merged together by cbind() to create one complete data frame data_All

  data_All<-cbind(data_Activity, data_Observations)  
  data_All<-cbind(data_Subject, data_All)
  

• Activity name for each activity, which has been loaded into data frame data_activity_labels in Step 1, is used to give proper description. Remove "_" from the activity label and convert it to lower case.

  data_activity_labels[,2]<-gsub("WALKING_UPSTAIRS","walkup", data_activity_labels[,2])
  data_activity_labels[,2]<-gsub("WALKING_DOWNSTAIRS","walkdown", data_activity_labels[,2])
  data_activity_labels[,2]<-tolower(gsub("_"," ", data_activity_labels[,2]))
  

• activity column of data frame data_All is converted to factor as

  data_All$activity<-as.factor(data_All$activity)
  

• Appropriate labels are assigned to the levels as

  setattr(data_All$activity, "levels", data_activity_labels[,2])
  

• Lower camel case is adopted for renaming the features for easy readability.

• Features are renamed to make it more descriptive by substituting Mean for -mean(), Std for -std(), timeDomain for t and frequencyDomain for f.

  names(data_All)<-gsub("-mean\\()","Mean", names(data_All)) 
  names(data_All)<-gsub("-std\\()","Std", names(data_All)) 
  names(data_All)<-gsub("^t","timeDomain", names(data_All)) 
  names(data_All)<-gsub("^f","frequencyDomain", names(data_All)) 
  

• There appeared to be an error in the naming of some features in the original dataset, for example fBodyBodyGyroJerkMag-mean() has Body repeated twice, hence all occurence of BodyBody is replaced by Body

  names(data_All)<-gsub("BodyBody","Body", names(data_All))           
  

• Assignment suggests to find variables pertaining to mean and standard deviations of various observations and produce a tidy dataset of the average of these variables for each combination of subject and activity.

• It is assumed that tidy data set is to be created for each variable extracted in previous step.

• library(reshape2) is required to perform this final step.

• melt() is used on data_All with id=c("subject", "activity") to create multiple rows of unique id-variable combinations.

  tidydata<-melt(as.data.frame(data_All), id=c("subject",  "activity"))
  

• Average of each variable for each activity and each subject is computed into final data frame tidydata using dcast()

  tidydata<-dcast(tidydata, subject + activity ~ variable, mean) 
  

• Rename all measure variables with prefix avg and maintaine the lower camel case convention.

  names(tidydata)<-gsub("^timeDomain","avgTimeDomain", names(tidydata)) 
  names(tidydata)<-gsub("^frequencyDomain","avgFrequencyDomain", names(tidydata)) 
  

• Write tidy data to a text file.

  write.table(tidydata, file = "tidyData.txt", row.names = FALSE, col.names = TRUE)