===============================================================================

The script initially downloaded the zip folder from the website ("UCI HAR Dataset.zip"), and saved it as "projectData.zip". Then "projectData.zip" will be "unzipped" to extract the sub-folders and files of the test and train datasets). The extracted folder that would be located in the working directory is called "UCI HAR Dataset". The contents of this folder is described below.
Finally, as cleanup, the copy of "projectData.zip" was deleted from the working directory by the unlink() command.

The "UCI HAR Dataset" folder contains a number of files and folders, but only the following files were used for processing into tidy data:

1. "activity labels": file containing a data frame of ActivityID's with the corresponding Activity name.
2. "features": a file containing the (561) descriptive names of data fields that can be used as column names for the test and train datasets (x_test and x_train).
3. "test" and "train" sub-FOLDERs containing the following files

• "subject_test" (data frame of SubjectID's) 2947 rows x 1 column
• "y_test" (data frame of ActivityID's) 2947 rows x 1 column
• "x_test" (data frame of 561 measuremnt variables) 2947 rows x 561 columns
• "subject_train" (data frame of SubjectID's) 7352 rows x 1 column
• "y_train" (data frame of ActivityID's) 7352 rows x 1 column
• "x_train" (data frame of 561 measuremnt variables) 7352 rows x 561 columns

*(Step 2 assumes that the "UCI HAR Dataset" folder containing the sub-folders and needed files, resides in the working directory)

The subject_test (2947 rows x 1 column) and y_test (2947 rows x 1 column) dataframes were read and column names (SubjectID and ActivityID, respectively) were attached.
The dataframes were then "cbind"-ed and the new, combined data frame was called subject.y_test (2947 rows x 2 columns).

Similarly, the subject_train (7352 rows x 1 column) and y_train (7352 rows x 1 column) dataframes were also read and column names (SubjectID and ActivityID, respectively) were attached.
The dataframes were then "cbind"-ed and the new, combined data frame was called subject.y_train (7352 rows x 2 columns) .

The subject.y_test DF will eventually be "cbind"-ed to the X_test DF, while the subject.y_train DF will eventually be "cbind"-ed to the X_train DF. In order to do this, the X_test and X_train data set files must first be read into R as a dataframe, and then given column names as explained in Steps 3 and 4.

Before reading the X_test and X_train data frames, the "features.txt" file was first read into a data frame and the vector of the feature names (i.e. column names of measured variables) was extracted (i.e., subsetted out) as a vector, features.vec.

The X_test (2947 rows x 561 columns) and X_train (7352 rows x 561 columns) dataset files were then read as dataframes in R, and column names were attached to each, using the previously obtained features.vec.

The new dataframe from the "cbind"ing of x_test and subject.y_test data frames was called "test_dataset"
The new dataframe from the "cbind"ing of x_train and subject.y_train data frames was called "train_dataset".

The test_dataset dataframe has 2847 rows x 561 columns.
The train_dataset dataframe has 7352 rows x 561 columns.

The resulting data frame from the row-binding (rbind) of the test_dataset and the train_dataset was called "Merged_Dataset" (10299 rows x 561 columns)

The activity labels data frame has two columns containing the "ActivityID" and "Activity" (i.e., activity name). This was used to properly match Activity (name) with ActivityID's in the Merged_Dataset, by the use of the "merge" command (in Step 7).

First, a merging "index" column was added into the Merged_Dataset, in order to preserve the row order of the Merged_Dataset dataframe when the merge function is applied. Inevitably, the merge function will destroy the row order of the merged dataframe (which would not be desirable in this case). The orginal row order can be restored by using the "merging index".
Once this was done, the "activity_labels" DF was merged with the Merged_Dataset DF. The merged dataframe was now the "updated" Merged_Dataset .

After merging, the Merged_Dataset DF would already contain the column of "Activity" (i.e., activity names).

The "dplyr" package was then used to properly arrange the columns of the Merged_Dataset, as follows:.

• The arrange() command was used to arrange the rows of the dataframe, according to the "merging index" (previously defined), thus preserving the original row order. Preserving the row order before and after merging is important to get the right dataframe.
• The select() command was used to get the needed columns for SubjectID, and all the remaining columns which were extracted by regular expressions containing "t", "f" and "a". The command also exluded the column for ActivityID by the "negative sign" in the argument, -ActivityID. The ActivityID column was excluded, because it was no longer necessary with the presence of the "Activity" (name) column (obtained from the merge operation described above).

The dplyr package in combination with regular expressions was used to extract the needed data containing means and standard deviations ("mean" and "std) of the measurement signals from the Merged_Dataset as follows:

• The select() command was used to get the columns for: (1) SubjectID, (2) Activity, (3) columns containing the regular expression ".[Mm]ean." , (4) columns containing the regular expression ".std.", and to exclude the columns containing "angle" and "Freq". As mentioned in my CodeBook, data columns for "Angle" and "Freq" were not included since they are calculated data, and not "measurement signals". Only measurement signals were taken in the making of the tidy data.

The data frame containing the extracted column names containing mean and standard deviation(s) (std) was called "Extracted_Dataset".

NOTE : As mentioned in my Code Book, only the mean and standard deviations for "measurement signals" were taken. (Please refer to my CodeBook for details).

This column names with mean and std's for "Angle" and "Freq" were not included, because by definition (in my Code Book), these were computed values, and not "measurement signals".

Thus the total number of columns was 68, broken down into 66 (mean and std) feature columns (of "measurement signals"), plus 1 column for SubjectID, plus 1 column for Activity. The total number of rows was 180, corresponding to combinations of 30 SubjectD's and 6 Activity's.

The "Activity" column of the Extracted_Dataset was also formalized as a factor with 6 levels of "activities"

The stringr package (particularly the str_replace and the str_replace_all commands) was used to convert the raw column names of the Extracted_Dataset, into those which comply with the requirements/conventions for naming of tidy data column names : Among these are: (1) human readable names, (2) no punctuations and/or symbols.
Also, the typo errors involving duplication of "Body" in "fBodyBodyAccJerkMag-std()", "fBodyBodyGyroMag-std()", and "ffBodyBodyGyroJerkMag-std()" were addressed.

The resulting column names after stringr manipulation were quite long, thus CamelCase had to be used in order to improve readability.

The dplyr package was again used to conveniently obtain the averages (i.e., the mean(s) of the column variables in the Extracted_Dataset, arranged according to Activity and SubjectID, and then saving the results into a new dataframe (called, PreFinal_TidyData) as follows:

• The group_by() command was first used to arrange the rows of the Extracted_Dataset according to SubjectID and Activity (i.e., activity name).
• The summarise_each(funs(mean)) command was then used to compute the means of the feature columns that were already previously arranged according to SubjectID and Activity.
• Finally, the arrange() command was used to make the primary ordering of the rows according to Activity

Since the PreFinal_TidyData now contained the average values (and not anymore the raw values) of the PreFinal_TidyData, the column names were ReLabelled to reflect that the values in this dataframe were already average values. The stringr package was used for this purpose, particularly the str_replace command.

The ReLabelled PreFinal_TidyData DF was finalized as the Final_TidyData, containing 180 rows (30 SubjectID''s x 6 Activity's) and 68 columns ( 1 for Activity + 1 for SubjectID + 66 for the averages of theextracted mean and std of measurement signals (pls refer to CodeBook and README))

The Final_TidyData was converted to a text file named, "Final_TidyData.txt", in compliance with the instructions of the Course Project.

NOTE: The saved text file can be read back into R by the following command:
TidyData <- read.table("Final_TidyData.txt", header=TRUE)