CourseProject_GettingCleaningData

---

Getting and Cleaning Data Course Project Version 1.0

Data is based on Human Activity Recognition Using Smartphones Dataset. It includes:

1. x_train
2. y_train
3. x_test
4. y_test
5. subject_train
6. subject_test
7. features.txt

Goal is to combine all the datasets above into one tidy dataset. The data in the x_test/train data contains sensor signal data (3 axis linear acceleration, gyroscope) for axial linear acceleration (x, y, and z) and angular velocity.

The dataset includes the following files:

• README.txt

• CodeBook.Rmd: a description of the data in the dataset (Alldata_tidy)

• run_analysis.R: an R script that creates the completed tidy dataset from the raw files. Steps:

  1. Read in the x_test, y_test, subject_test, x_train, y_train, subject_train data
  2. Read in the features.txt to obtain column names for step 1 (x_test, x_train)
  3. combine test and train data using rbind (x_train, x_test) [in this order]
  4. Determine the columns in the resulting data that are either mean or SD
  5. Subset the data (X data) so that it contains only mean/SD related columns
  6. Combine the subject data with the X data using cbind
  7. Read in the activity labels and then replace the values in y_test/train with the corresponding activity labels
  8. Combine the y_test/y_train data using rbind
  9. Combine X and Y data using rbind
  10. Rename the data columns in the resulting combined data to be descriptive
  11. Create Alldata_tidy by piping the dataset in step 10 into group_by (dplyr) using Subject and Activity columns as the grouping variables and then piping into summarize to calculate the mean of the other columns

• Alldata_tidy: the tidy dataset created from run_analysis.R. This data is tidy because:

  1. each variable forms a column,
  2. each observation forms a row, and
  3. each type of observational unit forms a table.

  An observation is a subject in an activity. The resulting dataset is long, rather than wide.