GettingandCleaningDataAssignment/run_analysis.R at master · cmsiu/GettingandCleaningDataAssignment · GitHub

43 lines (33 loc) · 1.65 KB

library(plyr)

##retrieve features.txt for getting required data column information

featuresFile <- read.table("features.txt")

##extract std() or mean() columns

extractCol <- sapply(featuresFile[,2], function(x) { if(grepl("std|mean[^Freq]", x)) "numeric" else "NULL"})

featureColName <- grep("std|mean[^Freq]", featuresFile[,2],value=TRUE)

##load activity labels

label <- read.table("activity_labels.txt", col.names=c("activity_id", "activity_name"))

##load test activities, dataset

teda <- read.table("test/y_test.txt", col.names="activity_id")

te_activity_name <- join(teda,label,by="activity_id",type="inner")[,"activity_name"]

##load test subject, dataset

teds <- read.table("test/subject_test.txt", col.names="subject_id")

tedd <- read.table("test/X_test.txt", colClasses=extractCol)

colnames(tedd) <- featureColName

testds <- cbind(teds,te_activity_name,tedd)

##load train activities, dataset

trda <- read.table("train/y_train.txt", col.names="activity_id")

tr_activity_name <- join(trda,label,by="activity_id",type="inner")[,"activity_name"]

##load train subject, dataset

trds <- read.table("train/subject_train.txt", col.names="subject_id")

trdd <- read.table("train/X_train.txt", colClasses=extractCol)

colnames(trdd) <- featureColName

trainds <- cbind(trds,tr_activity_name,trdd)

##combine test and train dataset

colnames(testds)[2] <- "activity_name"

colnames(trainds)[2] <- "activity_name"

ds <- rbind(testds,trainds)

##reshape dataframe to be a tidy dataset

molten = melt(ds, id=c("subject_id", "activity_name"), na.rm = TRUE)

tidyds <- dcast(molten, subject_id + activity_name ~ variable,mean)

##output tidyds

write.csv(tidyds, file = "tidy.txt", row.names=FALSE)

Expand file tree