Codebook.md

Codebook

This codebook breaks down algorithm of code run_analysis.R. The algorithm has the following key steps:

1. Downloading and unziping data from a given link to a folder UCI HAR Dataset. Prior downloading it checks whether the data has been already downloaded.
2. Using command read.table reading and writing the following data:

• features.txt to a data_names (chr [1:561]) that is lately used for naming columns of output data. The features selected for this database come from the accelerometer and gyroscope 3-axial raw signals tAcc-XYZ and tGyro-XYZ.
• activity_labels.txt to a data frame activity_label (6x2 df). This data frame links the class labels with their activity name.
• subject_test.txt and subject_train.txt to a data frames subject_test (2947x1 df) and subject_train (7352x1 df), respectively. Each row identifies the subject who performed the activity for each window sample. Its range is from 1 to 30.
• X_test.txt and X_train.txt to a data frames x_test (2947x561 df) and x_train (7352x561 df), respectively. Test and training sets' data is collected here. Names for columns are taken from data_names.
• Y_test.txt and Y_train.txt to a data frames y_test (2947x1 df) and y_train (7352x1 df), respectively. These data sets contain training activity number.

3. Using factor and sapply commands on activity_label, y_test and y_train dataframes, character labels are assignd to each observation of both test and train datasets creating new data frames called y_test_labeled (chr [1:2947]) and y_train_labeled (chr [1:7352]).
4. Using cbind command, above data frames are column binded to a subject_test andsubject_train data sets, respectively, creating data frame with subject number and activity done during each observation for both test and training sets. New data frames were named test_subj_act_label (2947x2 df) and train_subj_act_label (7352x2 df).
5. Using rbind command, x_test and x_train data frames were row binded together to form a new data frame called x_data (10299x561 df). The same wayy_data (10299x2 df) was created from y_test and y_train data frames.
6. Again using cbind command merged_data (10299x563 df) was obtained by binding y_data and x_data together.
7. Using select, only first two columns and columns that contain mean and std in the name were selected to form new data frame called mean_sd_data (10299x88 df)
8. using gsub command, namings of almost all columns were cleaned based on the specific pattern shown in the code.
9. Finally, the following code: mean_sd_data %>% group_by(Subject, Activity) %>% summarise_all(.funs = mean) was used in order to obtain new data fram average_data (180x88 df) containing means (using command summarise_all) and grouped by Subject and Activity (using command group_by)
10. Finally, average_data was written to a file average_data.txt using command write.table