quasiguru/samsung-activities

Getting and Cleaning Data Course Project README

Notes

• Uses R version 3.03
• Needs reshape2 library to be installed
• To read in the final tidy data set in R, use read.table("final.txt", header = TRUE, sep = "\t")

Original Data

• Full Experiment here
• Data also can found here
• Experiment [README](https://github.com/QuasiGuru/Samsung-Activities/blob/master/UCI HAR Dataset/README.txt)

File Structure

• README.MD -- this file

• CodeBook.MD -- data dictionary for cleaned up final data

• final.txt -- cleaned up tidy data set

• run_analysis.R -- file used to create tidy data set

• UCI HAR Dataset/README.txt -- description of original experiment and files

• UCI HAR Dataset/features_info.txt -- list of variables for features of original experiment

• UCI HAR Dataset/features.txt -- list of all features

• UCI HAR Dataset/activity_labels.txt -- list of activities

• UCI HAR Dataset/train/X_train.txt -- training set

• UCI HAR Dataset/train/y_train.txt -- training labels

• UCI HAR Dataset/train/subject_train.txt -- list of users for training set

• UCI HAR Dataset/test/X_train.txt -- testing set

• UCI HAR Dataset/test/y_train.txt -- testing labels

• UCI HAR Dataset/test/subject_train.txt -- list of users for testing set

Instrucions

Download The Original Data Set

• Download data from one of the links above and extract the files as shown above
• Place run_analysis.R in working directory (as suggested above)
• Run run_analysis.R

What run_analysis.R does

Step 1 -- Get The Features List

• Load the features (561 rows) from features.txt into a data frame
• Retrieve a list of the records that measure mean or standard deviation
• Clean up the feature name to make them cleaner and more descriptive

Step 2 -- Get The Activities List

• Load the features (6 rows) from activity_labels.txt into a data frame

Step 3 -- Get The Training List

• Load the features (7352 rows of 561 columns) from X_train.txt into a data frame
• Rename the column names to the names from the features list
• Choose only the columns based upon the list found in Step 1 that measure mean or standard deviation
• Load the subject ids (7352 rows) from subject_train.txt into a data frame and add that column to the training data frame
• Load the training activity ids (7352 rows) from y_train.txt into a data frame and add that column to the training data frame

Step 4 -- Get The Testing List

• Since we consider training and testing to be the same, the same process is used with the files X_test.txt, subject_test.txt, y_test.txt (all 2947 rows)

Step 5 -- Get The Full Training/Testing List

• Merge the training and testing data frames
• Add the appropriate activity name based upon the activity id
• Delete the activity id column (unneeded)

Step 6 -- Reshape The Data And Print It To A File

• Calculate the average values of the means/standard deviations for each actity and each subject
• Print the resulting data frame to file (tab delimited)