Correlation between sulfate and nitrate in pollution monitors

corr.R

corr <- function(directory, threshold = 0) {
   temp = list.files(path=paste("./",directory,sep=''),
                     pattern="*.csv", full.names = TRUE)
   data = do.call("rbind", lapply(temp, function(x) read.csv(x)))
   data$ID <- as.numeric(data$ID)

   data <- data[complete.cases(data), ]

   corrs <- as.numeric(c())
   for (i in unique(data$ID)){
      complete <- data[ which( data$ID == i ) , ]
      if (nrow(complete) > threshold)
         corrs <- append(corrs, cor(complete$sulfate, complete$nitrate))
   }
   corrs
}

pollutantmean <- function(directory, pollutant, id = 1:332) {

   temp = list.files(path=paste("./",directory,sep=''),
                     pattern="*.csv", full.names = TRUE)
   data = do.call("rbind", lapply(temp, function(x) read.csv(x)))
   data$ID <- as.numeric(data$ID)

   if (pollutant == 'sulfate'){
      sub <- subset(data, ID %in% id & !is.na(sulfate), select = sulfate)
      mean(sub$sulfate)
   } else if (pollutant == 'nitrate') {
      sub <- subset(data, ID %in% id & !is.na(nitrate), select = nitrate)
      mean(sub$nitrate)
   } else {
      stop("Err: unknow column")
   }

}

# counts complete readings for given set of monitors
complete <- function(directory, id = 1:332) {
   temp = list.files(path=paste("./",directory,sep=''),
                     pattern="*.csv", full.names = TRUE)
   data = do.call("rbind", lapply(temp, function(x) read.csv(x)))
   data$ID <- as.numeric(data$ID)

   data <- data[ which(complete.cases(data) & data$ID %in% id), c("ID")]
   data <- as.data.frame(data)
   colnames(data) <- c('occurences')

   result <- aggregate(data, by=list(data$occurences), FUN=length)
   colnames(result) <- c('id','nobs')

   result[order(match(result$id,id)), ]
}