TidyDataProject

Getting and Cleaning Data

Course Project

Mackenzie Wildman

1/22/17

Data source

• Human Activity Recognition Using Smartphones Dataset
• To run run_analysis.R, this data must be stored in the folder "UCI HAR Dataset" in the working directory, as downloaded from the assignment description page

Analysis

Data input

• Input measurement data from both trainX.txt and testX.txt into trainX and testX
• Input corresponding activities from y_train.txt and y_test.txt into trainy and testy
• Input corresponding subjects from subject_train.txt and subject_test.txt into trainsubject and testsubject
• Input key of labels for measurements from features.txt into featureslabels
• Input key of labels for activities from activity_labels.txt into activitylabels

Merge data horizontally

For both the train and test data sets, separately:

• Set measurement variable names in trainX and testX data frames from featureslabels
• Add activity numbers variable to both trainX and testX data frames from trainy and testy
• Convert activity numbers to activity labels in both trainX and testX data frames using activitylabels as key
• Add subject variable to both trainX and testX data frames from trainsubject and testsubject

Merge data vertically and make tiny data set

• Merge trainX and testX to make mergedData data frame
• Find measurement variable names containing mean or std
• Extract these variables along with subject and activity variables to make tinydata data frame

Create summary data set

• Melt tinydata data frame (reshape2 package) using subject and activity variables as id
• Reshape data frame using dcast, taking the mean of each measurement variable for each fixed subject, activity pair

To import summary data set

read.table("./summarydata.txt", header=TRUE)