The script run_analysis.R is designed to take two data sets from the same experiment, merge them, run some analysis, and tidy the data before writing it out to the file system. The script is split up into five sections, each of which will be discussed below.
The first step is to take two directories, test/ and train/, and merge their data into a single data set. In order to do that, we first read in all the file names from the test/ and train/ directories. Then we send those file paths into a function called mergeTwoFiles that takes two file paths, reads in the files, merges them out, and writes them out to the file system. Once that piece of the script completes, we're finished with step 1.
In this step, we intend to throw away all columns from the merged data aside from those indicating some sort of mean or standard deviation measurement. In order to do that, we read in the column names from the features.txt file, and then use a regular expression to extract only column names that contain the character group "mean" or "std", the indicators of mean and standard deviation that the dataset uses. We create a logical vector based on that regular expression, then subset the merged data by passing that logical vector into the columns argument of the subsetting operation. We're left with only the columns representing means or standard deviations from the merged data set.
In this step, we name the activities meaningfully, and merge them into the dataset. In order to do this, we create a character vector of human readable activities based on the activity_labels.txt file. Then we read in the file indicating the activities, and create a table where the numbered activities have been made human-readable by translating via the character vector we created. We then bind this column to the data.
Since the subjects are in a separate file from the measurements taken, for convenience we add the subjects to the table by reading in the subjects file, then binding the subjects to the data as a separate column.
In this step, we must name the columns meaningfully. In order to do that, we take the column names we found in step 2 and use string manipulation to remove any special characters from them and to make them more human readable. Once we've done that, we assign the column names to the data directly.
In the last step, we utilize the dplyr library, which makes summarizing data very easy in R. First we convert our data.frame into a tbl_df object. Then, using the group_by function, we add groups based on activity and subject to the data. From there, we can summarize the data based on these groups by taking the average of each measurement, and then write them out to the file system.
Once these steps have been performed, we have a file called tidy_data.txt in our directory which has our summarized, cleaned data!
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent