This project creates two "tidy" data sets from the UCI HAR data set

Dataset 1: meanstd.txt This data set contains the observations from values that are mean and standard deviations for each measurement

• columns chosen all had variables named mean() or std() for their respective features
• columns such as tBodyAccMean were not included becuase they were considered to be features themselves, not means or standard deviations of features

This dataset is tidy because: -Each row is an observation - a unique combination of subject, activity, type(test or training) -Each column is a variable:

• Identifier variables are subject.number; type, activity.name
• Measurement variable are the 66 mean and std values of the features/variables captured

Logical Approach - Each step below is a commented section of code

1. Read features.txt; and activity_labels.txt file

• These were read in first as individual files that were used to merge into final results
  • feature.txt contains the 561 feature/variable combinations that were in the measurement data
  • activity_labels.txt contains the names of the 6 activities

2. Identify feature numbers and feadturenamesof 561 containing "mean()" or "std()"

• select the sequence number of features containing mean( or std(; used later to subset the X data sets
• select the names associated with the selected feature; used later for column names

3. Modify featurenames to be used as for column names of new data frame

• edits the feature names for readable column headings
  • remove special characters ()
  • replace - with . as separater for X, Y, Z dimensions
  • replace t with time and f with frequency to be explicit
• formatting based on google R style with two exceptions for readability in this situation
  • . used to separate compound names
  • leading Captial letter retained, preservering data sets original designation of features and variables

4. Modify activty names to use as values (lower case)

• Consistent with goole R style

5. Read subject_train.txt; subject_test.txt; y_train.txt; y_test.txt

• This reads the data sets that are "clubbed" to the main data (subject number and activity number)

6. Read X_train.txt and X_test.txt - columns with mean() and std()

• This reads the main data sets, limitng the columns stored to those identified as mean( or std( in step 2
• This uses the associated names in step 2, with edits in step 3 as column names

7. Add type column to xtrain and xtest

• Clubs a column for type, with fixed value of test for the test data set and train for the train data set

8. Add activity to xtest and xtrain

• Clubs a column for activity number from the Y data sets read in step 5.

9. Add subject to xtest and xtrain

• Clubs a column for subject number from the subject data sets read in step 5.

10. Replace activity number with Activity Name

• Merges the Activity name from data read in step 1.; and deletes activity number

11. Combine test and training data and reorder columns to put all mean and std variable to right

• Combines the rows from test and train
• reorders columns so that all idnetifier variables are to the left and measurement variables to the write

12. Create summarymeanstd.txt - summmary data frame

• Dataframe remains "tidy"
  • each row is now an observation of mean of subject and activity (30 subjects x 6 activities yields 180 rows)
  • each column remains the identifier variables, and the 66 features/variables of mean and std