I used the packages sqdf,plyr

First I read all the files, giving them the names rawtest,rawtrain etc..

then I add a new column for subject data on both rawtest and rawtrain then I add a new column for category data( wether it's test or train data), on both rawtest and rawtrain then I add a new column for activities data( 1:6), on both rawtest and rawtrain

then I add a new column to features data for labels subject, activity, and category)

Finally I merged the dataframe of features with the rawdata so that each coulmn has a name and then I merged test and train data using rbind

in the end I subset the "new" dataframe to find only column with mean

But since it takes also meanFreq() value I had to filter the data excluding those that were Freq

and in the end I made another dataframe only for standard deviation

And to create the new tidy data sets I use the aggregate function, while to write the document in txt the write.table function