###run_analysis.R

1. run_analysis.R is run against a provided data set and creates another data set that is clean and tidy.

2. run_analysis.R requires the installation of the "reshape2" library

3. run_analysis.R assumes either 1) that the original data set is moved into the working directory or 2) that the working directory is set to the folder that contains the original data set.

4. The codebook entitled "Movement_Data_Dictionary.pdf" provides information about the original data set that is used by run_analysis.R and how the data is transformed.

5. Here is a line-by-line commentary on how run_analysis.R transforms the original data set:

   Test data is imported into a table test_data

   test_data = read.table("test/x_test.txt")
   

   Train data is imported into a table train_data

   train_data = read.table("train/x_train.txt")
   

   test_data and train_data tables are combined into total_data

   total_data <- rbind(test_data,train_data)
   

   Column headings for the data are imported and converted to text

   labels = read.table("features.txt")
   labels_data <- as.character(labels[,2])
   

   Column headings are applied to the data table total_data

   colnames(total_data) <- labels_data
   

   A vector is created that consists of the column numbers of the column headings that contain text of "mean", "std", or "Mean".

   e <- sort(c(grep("mean",labels_data),grep("std",labels_data),grep("Mean",labels_data)))
   

   total_data is reduced to columns identified in previous step and assigned to total_data2

   total_data2 <- total_data[,e]
   

   Activity codes for test data are imported

   test_labels = read.table("test/y_test.txt")
   

   Activity codes for train data are imported

   train_labels = read.table("train/y_train.txt")
   

   Activity codes for test and train data are combined

   total_labels <- rbind(test_labels,train_labels)
   

   Subject codes for test data are imported

   test_subjects = read.table("test/subject_test.txt")
   

   Subject codes for train data are imported

   train_subjects = read.table("train/subject_train.txt")
   

   Subject codes for test and train are combined

   total_subjects <- rbind(test_subjects,train_subjects)
   

   Activity and Subject codes are combined with data

   total_data3 <- cbind(total_subjects,total_labels,total_data2)
   

   Column names are applied to the data table

   colnames(total_data3) <- c("Subject_ID","Activity",labels_data[e])
   

   Activity labels that correlate to the Activity codes are imported and converted to text

   activities = read.table("activity_labels.txt")
   activities_col2 <- as.character(activities[,2])
   

   Activity codes are converted to Activity labels

   for (i in 1:nrow(total_data3)) {
     total_data3[i,2] <- (if (total_data3[i,2] == "1") activities_col2[1] else if (total_data3[i,2] == "2") activities_col2[2] else if (total_data3[i,2] == "3") activities_col2[3] else if (total_data3[i,2] == "4") activities_col2[4] else if (total_data3[i,2] == "5") activities_col2[5] else if (total_data3[i,2] == "6") activities_col2[6])
   }
   

   Data table is "melted" into id and variable categories for use in the next step

   total_data4 <- melt(total_data3,id.vars=c("Subject_ID","Activity"))
   

   A table is created showing the mean of each variable, broken down by Subject and Activity

   total_data5 <- dcast(total_data4,Subject_ID + Activity ~ variable,mean)