Skip to content

Latest commit

 

History

History
29 lines (21 loc) · 1.84 KB

README.md

File metadata and controls

29 lines (21 loc) · 1.84 KB

Getting-and-Cleaning-Data-Project

In this README file, I will go through the code, in order to demonstrate how I imeplented the parsing script. Of course you can always check the comments in the script, but they should be complementary to each other.

Step 1 - Set the Workspace

This should be adapted to someone's own computer setting (where the data is located). We put all the txt files directly under the workspace folder.

Step 2 - Select features from the 561 feature spaces

  • We first read the features list as a dataframe,
  • then we perform string operations to select those features that contain "mean" or "std".
  • After that, we have the features that need to be extracted (79 out of 561), save these features as a variable called "featureToExtract"
  • Then change the original feature names by removing brackets and changing "-" to "."

Step 3 - Parse the training data as well as the testing data

  • Read subject list
  • Read all feature variables, but we subset only the variables that are in "featureToExtract"
  • Change variable names to that in the "featureToExtract"
  • Read activities list (y_test or y_train)
  • subject list, activity list and the feature list should be of the same length, so we can bind them together using cbind()
  • We change the non-feature varabiles to "subject","activity" and "data.type". This dataset is the tidy1.csv

Step 4 - Take average per subject per activity (required by the 5th requirement)

  • Use a convinient function aggregate() to group the variables in one line. This dataset is the tidy2.csv

NOTE: Running the script

Simply select all the content of the script run_analysis.R in Rstudio, and click run, you will get the two files. However, you need to change the workspace to the corresponding data folder on your computer.