Cleaning.Data_Ass3

Step 1

Should first use loop to build path for reading text from subfolders

set <- c("test","train")
xysub <- c("X", "y", "subject")

 for (i in 1:2){
    ** for (j in 1:3){
        url <- paste("./",set[i],"/",xysub[j],"_",set[i],".txt",sep="")
        # use function assign to smooth the loop
        assign(paste(tolower(xysub[j]),"_",set[i],sep=""),read.table(url,header=FALSE,sep=""))
    }
}

Use rbind() to mergr Train set and Test Set

all_x <- rbind(x_test, x_train)
all_y <- rbind(y_test, y_train)
all_sub <- rbind(subject_test, subject_train)

Step 2

Calculate Mean and sd of each variable after merging

The apply() would simply represent a list of with means and sd of 561 variable

allx_mean <- apply(all_x,2,mean)
allx_sd <- apply(all_x, 2, sd)

Step 3

extract names from activity_labels.txt and apply them put them in dataframe for Y

Use factors to match naems

act_name <- read.table("./activity_labels.txt")
names_Y <- factor(all_y$V1,labels = as.vector(act_name$V2))
y.name <- as.vector(act_name$V2)
all_y_with_names <- data.frame(names_Y)

Step 4

similarly, extract names from features.txt

names_561 <- read.table("./features.txt")
x2_all <- all_x
var.name <- as.vector(names_561$V2)
rm(names_Y, names_561)
colnames(x2_all) <- var.name
colnames(all_sub) <- c("Subject")
colnames(all_y_with_names) <- c("Act")

Answer of step 4 after putting the column name

df <- cbind(all_sub,all_y_with_names,x2_all)

Step 5

loop to subset data into smaller dataset for each subject and each activity

i.e. should loops 180 times for 30 subject x 6 activity

Use apply() to return col names with a list of 561 variables

Use rbind() to build a new dataframe each time when loop is about to finished

Finally, put columns of subject and activity back ino dataframe to export data

df_0 <- data.frame()

for (i in 1:30){
    df_p <- df[df$Subject==i,]
    for (j in 1:6){
        df_sum <- 0
        df_q <- df_p[df_p$Act == y.name[j],]
    df_sum <- apply(df_q[,3:563], 2, mean)
    df_0 <- rbind(df_0,df_sum)
    }
}

test <- rep(c(1:30), each = 6)
test2 <- rep(y.name, length.out = 180)
df_q5 = cbind(test, test2, df_0)
rm(df_0,test,test2)
colnames(df_q5) = c("Subject", "Act",var.name)

export the data.frame in summary.txt without row.names

write.table(df_q5,"./summary.txt",row.name=FALSE)