GCDCourseProject

run_analysis.R

DATA_ROOT < 'UCI HAR Dataset'
curr.sep <- .Platform[['file.sep']]

read.commons <- function(){

  act.labels <<- read.table(paste(DATA_ROOT
                                ,'activity_labels.txt'
                                , sep=curr.sep)
    ,sep=' '
    ,col.names=c('index','name'))

  col.names <- read.table(paste(DATA_ROOT
                                ,'features.txt'
                                , sep=curr.sep)
    ,stringsAsFactors=F
    ,sep=' ')[[2]]

  selected.cols     <<- grepl('-mean()|-std()',col.names)
  # pretty.col.names  <<- gsub('([[:lower:]])([[:upper:]])'
  #                           , '\\1 \\2'
  #                           , gsub( '[.-]|\\(\\)'
  #                                 , ' '
  #                                 , col.names[selected.cols]))

  pretty.col.names  <<- col.names[selected.cols]
}

read.datasets <- function(dataset.name){

  y.data <- read.table(paste(DATA_ROOT
                           , dataset.name
                           , paste0('y_',dataset.name,'.txt')
                           , sep=curr.sep
                            )
                      )[[1]]

  y.data <- act.labels[y.data,2]

  subj.data <- read.table(paste(DATA_ROOT
                             , dataset.name
                             , paste0('subject_',dataset.name,'.txt')
                             , sep=curr.sep
                              )
                        )[[1]]

  x.data <- read.table(text=gsub( '[[:space:]]+'
                                , ';'
                                , sub( '^[[:space:]]+'
                                     , ''
                                     , readLines(paste( DATA_ROOT
                                                      , dataset.name
                                                      , paste0('X_',dataset.name,'.txt')
                                                      , sep=curr.sep)
                                                      )
                                                )
                                )
                     , sep=';'
                      )[,selected.cols]

  cbind(as.factor(subj.data),y.data,x.data)
}

read.commons()

result.df <- rbind(read.datasets('test'),read.datasets('train'))
names(result.df) <- c('subject','activity',pretty.col.names)

agg <- aggregate(. ~ subject + activity, data = result.df, mean)

write.table(agg,paste(DATA_ROOT,'output.txt',sep=curr.sep),sep=';',col.names=names(agg))