Skip to content

CodeBook.md

Latest commit

 

History

History
117 lines (103 loc) · 10.5 KB

CodeBook.md

File metadata and controls

117 lines (103 loc) · 10.5 KB

The CodeBook

Introduction

This data is used for "Getting and Cleaning Data" from Coursera. The original data can be downloaded from http://archive.ics.uci.edu/ml/datasets/Human+Activity+Recognition+Using+Smartphones.

This data comes from experiments carried out by Jorge L. Reyes-Ortiz, Davide Anguita, Alessandro Ghio, Luca Oneto from Smartlab - Non Linear Complex Systems Laboratory. They aquired measurements from inertial sensors (accelerometer, gyroscope), which were built into Samsung Galaxy S II smartphones used by 30 volunteers. After filtering some features were computed. For more information please read README.txt from http://archive.ics.uci.edu/ml/machine-learning-databases/00240/UCI%20HAR%20Dataset.zip.

Description

In this case both the training and the testing data were merged and only several features were extracted (these, which were described as a mean, a standard deviation or a frequency mean).

If the feature name starts with "time", it was computed using a time domain analysis, similarily if the feature name starts with "frequency", it was computed using a frequency domain analysis. There are two groups of data: a body and a gravity. These two groups were extracted from the raw data using a Butterworth filter. As mentioned above, two inertial sensors were used: an accelerometer and a gyroscope. Several variables were computed: X, Y, Z stand for direction, Magnitude were computed using the Euclidean norm and the body linear acceleration and angular velocity were derived in time to obtain Jerk signals. Frequency domain features were computed using Fast Fourier Transform.

Apart from the "activity" which is a string and the "subject" which is a number from 1 to 30, all features are normalized and bounded within [-1,1].

Feature list

There are following features in the dataset:

Feature Domain Device Group Value type Comment
activity NA NA NA NA an activity that the volunteer performed
subject NA NA NA NA an identifier of the subject who carried out the experiment (from 1 to 30)
timeBodyAccelerometerMeanX time accelerometer body mean x direction
timeBodyAccelerometerMeanY time accelerometer body mean y direction
timeBodyAccelerometerMeanZ time accelerometer body mean z direction
timeBodyAccelerometerStandardDeviationX time accelerometer body standard deviation x direction
timeBodyAccelerometerStandardDeviationY time accelerometer body standard deviation y direction
timeBodyAccelerometerStandardDeviationZ time accelerometer body standard deviation z direction
timeGravityAccelerometerMeanX time accelerometer gravity mean x direction
timeGravityAccelerometerMeanY time accelerometer gravity mean y direction
timeGravityAccelerometerMeanZ time accelerometer gravity mean z direction
timeGravityAccelerometerStandardDeviationX time accelerometer gravity standard deviation x direction
timeGravityAccelerometerStandardDeviationY time accelerometer gravity standard deviation y direction
timeGravityAccelerometerStandardDeviationZ time accelerometer gravity standard deviation z direction
timeBodyAccelerometerJerkMeanX time accelerometer body mean x direction (jerk)
timeBodyAccelerometerJerkMeanY time accelerometer body mean y direction (jerk)
timeBodyAccelerometerJerkMeanZ time accelerometer body mean z direction (jerk)
timeBodyAccelerometerJerkStandardDeviationX time accelerometer body standard deviation x direction (jerk)
timeBodyAccelerometerJerkStandardDeviationY time accelerometer body standard deviation y direction (jerk)
timeBodyAccelerometerJerkStandardDeviationZ time accelerometer body standard deviation z direction (jerk)
timeBodyGyroscopeMeanX time gyroscope body mean x direction
timeBodyGyroscopeMeanY time gyroscope body mean y direction
timeBodyGyroscopeMeanZ time gyroscope body mean z direction
timeBodyGyroscopeStandardDeviationX time gyroscope body standard deviation x direction
timeBodyGyroscopeStandardDeviationY time gyroscope body standard deviation y direction
timeBodyGyroscopeStandardDeviationZ time gyroscope body standard deviation z direction
timeBodyGyroscopeJerkMeanX time gyroscope body mean x direction (jerk)
timeBodyGyroscopeJerkMeanY time gyroscope body mean y direction (jerk)
timeBodyGyroscopeJerkMeanZ time gyroscope body mean z direction (jerk)
timeBodyGyroscopeJerkStandardDeviationX time gyroscope body standard deviation x direction (jerk)
timeBodyGyroscopeJerkStandardDeviationY time gyroscope body standard deviation y direction (jerk)
timeBodyGyroscopeJerkStandardDeviationZ time gyroscope body standard deviation z direction (jerk)
timeBodyAccelerometerMagnitudeMean time accelerometer body mean magnitude
timeBodyAccelerometerMagnitudeStandardDeviation time accelerometer body standard deviation magnitude
timeGravityAccelerometerMagnitudeMean time accelerometer gravity mean magnitude
timeGravityAccelerometerMagnitudeStandardDeviation time accelerometer gravity standard deviation magnitude
timeBodyAccelerometerJerkMagnitudeMean time accelerometer body mean magnitude (jerk)
timeBodyAccelerometerJerkMagnitudeStandardDeviation time accelerometer body standard deviation magnitude (jerk)
timeBodyGyroscopeMagnitudeMean time gyroscope body mean magnitude
timeBodyGyroscopeMagnitudeStandardDeviation time gyroscope body standard deviation magnitude
timeBodyGyroscopeJerkMagnitudeMean time gyroscope body mean magnitude (jerk)
timeBodyGyroscopeJerkMagnitudeStandardDeviation time gyroscope body standard deviation magnitude (jerk)
frequencyBodyAccelerometerMeanX frequency accelerometer body mean x direction
frequencyBodyAccelerometerMeanY frequency accelerometer body mean y direction
frequencyBodyAccelerometerMeanZ frequency accelerometer body mean z direction
frequencyBodyAccelerometerStandardDeviationX frequency accelerometer body standard deviation x direction
frequencyBodyAccelerometerStandardDeviationY frequency accelerometer body standard deviation y direction
frequencyBodyAccelerometerStandardDeviationZ frequency accelerometer body standard deviation z direction
frequencyBodyAccelerometerMeanFrequencyX frequency accelerometer body mean frequency x direction
frequencyBodyAccelerometerMeanFrequencyY frequency accelerometer body mean frequency y direction
frequencyBodyAccelerometerMeanFrequencyZ frequency accelerometer body mean frequency z direction
frequencyBodyAccelerometerJerkMeanX frequency accelerometer body mean x direction (jerk)
frequencyBodyAccelerometerJerkMeanY frequency accelerometer body mean y direction (jerk)
frequencyBodyAccelerometerJerkMeanZ frequency accelerometer body mean z direction (jerk)
frequencyBodyAccelerometerJerkStandardDeviationX frequency accelerometer body standard deviation x direction (jerk)
frequencyBodyAccelerometerJerkStandardDeviationY frequency accelerometer body standard deviation y direction (jerk)
frequencyBodyAccelerometerJerkStandardDeviationZ frequency accelerometer body standard deviation z direction (jerk)
frequencyBodyAccelerometerJerkMeanFrequencyX frequency accelerometer body mean frequency x direction (jerk)
frequencyBodyAccelerometerJerkMeanFrequencyY frequency accelerometer body mean frequency y direction (jerk)
frequencyBodyAccelerometerJerkMeanFrequencyZ frequency accelerometer body mean frequency z direction (jerk)
frequencyBodyGyroscopeMeanX frequency gyroscope body mean x direction
frequencyBodyGyroscopeMeanY frequency gyroscope body mean y direction
frequencyBodyGyroscopeMeanZ frequency gyroscope body mean z direction
frequencyBodyGyroscopeStandardDeviationX frequency gyroscope body standard deviation x direction
frequencyBodyGyroscopeStandardDeviationY frequency gyroscope body standard deviation y direction
frequencyBodyGyroscopeStandardDeviationZ frequency gyroscope body standard deviation z direction
frequencyBodyGyroscopeMeanFrequencyX frequency gyroscope body mean frequency x direction
frequencyBodyGyroscopeMeanFrequencyY frequency gyroscope body mean frequency y direction
frequencyBodyGyroscopeMeanFrequencyZ frequency gyroscope body mean frequency z direction
frequencyBodyAccelerometerMagnitudeMean frequency accelerometer body mean magnitude
frequencyBodyAccelerometerMagnitudeStandardDeviation frequency accelerometer body standard deviation magnitude
frequencyBodyAccelerometerMagnitudeMeanFrequency frequency accelerometer body mean frequency magnitude
frequencyBodyAccelerometerJerkMagnitudeMean frequency accelerometer body mean magnitude (jerk)
frequencyBodyAccelerometerJerkMagnitudeStandardDeviation frequency accelerometer body standard deviation magnitude (jerk)
frequencyBodyAccelerometerJerkMagnitudeMeanFrequency frequency accelerometer body mean frequency magnitude (jerk)
frequencyBodyGyroscopeMagnitudeMean frequency gyroscope body mean magnitude
frequencyBodyGyroscopeMagnitudeStandardDeviation frequency gyroscope body standard deviation magnitude
frequencyBodyGyroscopeMagnitudeMeanFrequency frequency gyroscope body mean frequency magnitude
frequencyBodyGyroscopeJerkMagnitudeMean frequency gyroscope body mean magnitude (jerk)
frequencyBodyGyroscopeJerkMagnitudeStandardDeviation frequency gyroscope body standard deviation magnitude (jerk)
frequencyBodyGyroscopeJerkMagnitudeMeanFrequency frequency gyroscope body mean frequency magnitude (jerk)

Processing path

  1. Merging data from both test and train datasets.
  2. Extracting features on the mean and the standard deviation (without angle() variables).
  3. Using descriptive labels for each variable.
  4. Using descriptive values for the "activity" variable.
  5. Computing the average of each variable for every pair (activity, subject) that occurs in the dataset.

Files

  • cleaned/cleaned_data.txt - the merged, extracted and described dataset that contains all features described above (after step 4.)
  • cleaned/averages.txt - the list of averages for each variable for each pair ("activity", "subject") that occurs in the cleaned_data.txt file