Cleaning_Getting_Assignment_coursera

This repository is for the Cleaning and Getting Data Assignment of coursera R programming specialization. R Script uses dplyr and needs to be installed

Working Directory

The following R code was designed to working inside a directory that has just another folder with the unzip files from the Assignments as they are and with no other modifications.

• It should look like this:

• UCI HAR Dataset
  • test
    • Invertial Signals
    • subject_test.txt
    • X_test.txt
    • Y_test.txt
  • train
    • Invertial Signals
    • subject_test.txt
    • X_train.txt
    • Y_train.txt
  • activity_labels.txt
  • features.txt
  • features_info.txt

Step 1 (Lines 2-30)

• Data is read from the different directories of the folder UCI HAR DATASET
  • Test datasets are processed
  • Train datasets are processed
  • Features dataset is processed and cleaned in order to avoid special characters such as (), - and ,.
  • Activity dataset is processed and all words are set to lower case.
  • Subjects datasets are processed

Step 2

• In order to match the activity accordingly i use the merge function

Step 3 (Lines 36-40) (Optional)

• Just in case someone would wanted to add a "Type" variable to distinguish between train and test datasets. You would need to add the type_train and the type_test variables in the cbind in lines 42 and 43.

Step 4

• Datasets are bind together depending if they are train or test datasets
• Later rbind is used since they have the same number of columns.

Step 5

• grep is used to get the index of which columns we should keep. Either if they are have the words mean or std.
• Since we also have meanFreq and is not a mean variable per se I have decided to omit it.
• final_dataset is created which is grouped by activity and subject and has the means of all the variables previously subsetted.
• A new txt file is created under the same name.