The R script works properly if the UCI HAR Dataset directory is next to it. This a condition of reading the input files.
The code consists of 4 parts:
- Merging the training and test sets
- Naming the columns appropriately and exracting only the mean and standard deciation measurements
- Including descriptive value names to the target attribute
- Creating the final aggregated data set
The script reads all 3 training files from their directory and with the cbind() function. Then it does the same for the 3 test files. Finally, it puts all previous data to the allData dataframe with an rbind().
The second part first reads the feature names from the features.txt file into the nameVector vector. We name the columns accordingly to them (except for the first and second column, whose name we know). Thus, we get interpretable names for our columns.
Using the grepl function, we are able to find which column we need finally. In the neededData dataframe we only put these needed columns.
The third part reads the activity names from activity_labels.txt into the actLabes dataframe. We merge this data frame with the neededData so for each line we get the tidy name of each activity. We arrange that the new column that has these values will have the activity_label name.
With the aggregate function, the script calculates the mean value for reach activity and each subject. For that, both activity_label and subject need to be a factor variable. We finally write the new tidy data set into the independent_tidy.txt file.