Skip to content

CodeBook.md

Latest commit

 

History

History
82 lines (61 loc) · 6.35 KB

CodeBook.md

File metadata and controls

82 lines (61 loc) · 6.35 KB

CodeBook for Getting and Cleaning Data - Course Project

Project Description

Data used by this analysis comes from accelerometer and gyroscope 3-axial raw signals recorded via Samsung Galaxy S II smartphones worn by 30 subjects performing a number of activities.

For a fuller understanding of the original dataset, please see features_info.txt in the referenced data.

Raw Data Collection

Per the README.txt file available from above source:

The experiments have been carried out with a group of 30 volunteers within an age bracket of 19-48 years. Each person performed six activities (WALKING, WALKING_UPSTAIRS, WALKING_DOWNSTAIRS, SITTING, STANDING, LAYING) wearing a smartphone (Samsung Galaxy S II) on the waist. Using its embedded accelerometer and gyroscope, we captured 3-axial linear acceleration and 3-axial angular velocity at a constant rate of 50Hz. The experiments have been video-recorded to label the data manually. The obtained dataset has been randomly partitioned into two sets, where 70% of the volunteers was selected for generating the training data and 30% the test data.

The sensor signals (accelerometer and gyroscope) were pre-processed by applying noise filters and then sampled in fixed-width sliding windows of 2.56 sec and 50% overlap (128 readings/window). The sensor acceleration signal, which has gravitational and body motion components, was separated using a Butterworth low-pass filter into body acceleration and gravity. The gravitational force is assumed to have only low frequency components, therefore a filter with 0.3 Hz cutoff frequency was used. From each window, a vector of features was obtained by calculating variables from the time and frequency domain. See 'features_info.txt' for more details.

About the test/train data files

The test and train data sets each contain three separate text files.

  1. X_(test|train) - Rows contain feature measurements.
  2. y_(test|train) - Contain activity IDs corresponding to each row in X tables; readable values of representing activities are stored in activity_labels,txt file.
  3. subject_(test|train) - Contain identifier for subject of each measurement/activity in the X and y files.

##Data Loading, Transformation, and Output

###Step 1 - Merge the test/training data to create single data set

  1. Data from each test/training file is loaded to individual data frames. Column headers for the y and subject data frames are changed to activityLabel and subjectId, respectively, for readability.
    • X_(test|train) -> df_(test|train)_X
    • y_(test|train) -> df_(test|train)_y
    • subject_(test|train) -> df_(test|train)_subject
  2. Data frames for test are combined using bind_cols() into df_test_all, followed by same action for train files into df_train_all.
  3. df_test_all and df_train_all are combined into df_combined using bind_rows().

###Step 2 - Extract mean/standard deviation variables only

  1. Features are loaded from features.txt into data frame df_features, filtering for those that contain the mean or std using the combined filter() and grep1() commands.
  2. mutate() is used to rename values in first column so they match the variable names in the combined data frame.
  3. Vector is created from df_features using select() and collect() that contains only mean/standard deviation related variables. subjectId and activityLabel are added to beginning of vector to allow for 1-to-1 match when doing select on df_combined.
  4. select command is used to extract relevant columns from df_combined; results are written to new data frame df_combined_select.

###Step 3 - Change activity names to descriptive values

  1. activity_labels.txt is loaded into df_activity_labels. Columns are renamed ID and activity for match function usage.
  2. activityLabel factor in df_combined_select is renamed using match() between df_activity_labels$ID and the df_combined_select$activityLabel.

###Step 4 - Label variables with descriptive variable names

  1. Vector is created containing from df_features using select() and collect() that contains only mean/standard deviation related variables. subjectId and activityLabel are added to beginning of vector so they will be placed correctly when df_combined_select's column names are updated.
  2. Vector values are written to colnames(df_combined_select).
  3. gsub() is used to improve readability of variable names.
    names(df_combined_select)<-gsub("^t", "time", names(df_combined_select))
    names(df_combined_select)<-gsub("^f", "frequency", names(df_combined_select))
    names(df_combined_select)<-gsub("Acc", "Accelerometer", names(df_combined_select))
    names(df_combined_select)<-gsub("Gyro", "Gyroscope", names(df_combined_select))
    names(df_combined_select)<-gsub("Mag", "Magnitude", names(df_combined_select))
    names(df_combined_select)<-gsub("BodyBody", "Body", names(df_combined_select))
    names(df_combined_select)<-gsub("\\(\\)", "", names(df_combined_select))

###Step 5 - Group data, get mean for each variable, and output to file

  1. group_by(subjectId,activityLabel) is used on df_combined_select.
  2. Above is piped to summarize_each(funs(mean)) to determine mean of each function when grouped by subject and activity, which is then written to new data frame df_combined_means.
  3. write.table() generates meansBySubjectAndActivity.txt` in working directory.

##Description of tidy data set

Tidy data set is 180 rows by 68 columns, each row containing:

  • subjectId - Numeric ID representing an individual subject
  • activityLabel - one of six activities performed by each subject; valid activities are:
    • WALKING
    • WALKING_UPSTAIRS
    • WALKING_DOWNSTAIRS
    • SITTING
    • STANDING
    • LAYING
  • 66 variables (Numeric) measuring representing the overall average of either the mean (variable names containing mean) or standard deviation (variable names containg std) for the specified time domain signal or frequency domain signal when grouped by subjectId and activityLabel. Variables ending with "-XYZ" represent 3-axial signals in the X, Y and Z directions.
  • Explanation of features can be found in features_info.txt from source. Names in tidy data set have been expanded per Step 4 above and parentheses have been removed for clarity.