deepakagg / coursera-getting-and-cleaning-data

Getting and cleaning data

For creating a tidy data set of wearable computing data originally from http://archive.ics.uci.edu/ml/datasets/Human+Activity+Recognition+Using+Smartphones

Files in this repo

• README.md -- you are reading it right now
• CodeBook.md -- codebook describing variables, the data and transformations
• run_analysis.R -- actual R code

run_analysis.R goals

You should create one R script called run_analysis.R that does the following:

1. Merges the training and the test sets to create one data set.
2. Extracts only the measurements on the mean and standard deviation for each measurement.
3. Uses descriptive activity names to name the activities in the data set
4. Appropriately labels the data set with descriptive activity names.
5. Creates a second, independent tidy data set with the average of each variable for each activity and each subject.

It should run in a folder of the Samsung data (the zip had this folder: UCI HAR Dataset) The script assumes it has in it's working directory the following files and folders:

• activity_labels.txt
• features.txt
• test/
• train/

The output is created in working directory with the name of tidy2.txt

Note: the R script is built to run without including any libraries for the purpose of this course.

run_analysis.R walkthrough

It follows the goals step by step.

• Step 1:

  • Read all the test and training files: y_test.txt, subject_test.txt and X_test.txt.
  • Combine the files to a data frame in the form of subjects, labels, the rest of the data.

• Step 2:

  • Read the features from features.txt and filter it to only leave features that are either means ("mean()") or standard deviations ("std()"). The reason for leaving out meanFreq() is that the goal for this step is to only include means and standard deviations of measurements, of which meanFreq() is neither.
  • A new data frame is then created that includes subjects, labels and the described features.

• Step 3:

  • Read the activity labels from activity_labels.txt and replace the numbers with the text.

• Step 4:

  • Make a column list (includig "subjects" and "label" at the start)
  • Tidy-fy the list by removing all non-alphanumeric characters and converting the result to lowercase
  • Apply the now-good-columnnames to the data frame

• Step 5:

  • Create a new data frame by finding the mean for each combination of subject and label. It's done by aggregate() function

• Final step:

  • Write the new tidy set into a text file called tidy2.txt, formatted similarly to the original files.