first commit

run_analysis.R

@@ -0,0 +1,39 @@
+require(pylr)
+# Read train and test data, 
+data.filename <- c("data//train/X_train.txt", "data//test//X_test.txt")
+data <- ldply(data.filename, read.table)
+
+# Read labels, assign to colnames(data) to finish Step 1
+label <- read.table("data/features.txt")[, 2]
+colnames(data) <- label
+
+
+# Extract activity id and corresponding label
+activity.filename = c("data//train//y_train.txt", "data//test//y_test.txt")
+activity <- ldply(activity.filename, read.table)
+colnames(activity) <- c("activity_id")
+
+activity.label <- read.table("data//activity_labels.txt", col.names=c("activity_id", "activity_label"))
+activity <- merge(activity, activity.label)
+# add it to data
+data$activity_label <- activity$activity_label
+
+# Similarly, extract subject id
+subject.filename = c("data//train//subject_train.txt", "data//test//subject_test.txt")
+subject <- ldply(subject.filename, read.table)
+
+# Add it to data, thus finishing Step 3 and 4
+data$subject_id <- subject[, 1]
+
+# Extracting those colnames which contains "std()" and "mean()" 
+# using customized filter function
+# thus finish Step 2
+contains_mean_std <- function (s) {
+  return(grepl("-std()", s) || grepl("-mean()", s))
+}
+
+mean_std_data <- data[, sapply(label, contains_mean_std)]
+
+# Calculate the mean for each (subject, activity) pair
+result <- ddply(mean_std_data, .(subject_id, activity_label), numcolwise(mean))
+print(result)