data/scripts/augment_meteo.R

   1 meteo_df = read.csv("meteo.csv")
   2
   3 meteo_df$Season = 0
   4 meteo_df$Week = 0
   5 meteo_df$Pollution = -1
   6
   7 #Need to load and aggregate PM10 by days
   8 pm10_df = read.csv("pm10_mesures.csv")
   9
  10 line_number = 1 #line number in pm10 file
  11 for (i in 1:(nrow(meteo_df)))
  12 {
  13     pm10s = c()
  14     repeat {
  15     {
  16         pm10s = c(pm10s, pm10_df[line_number,2])
  17         line_number = line_number + 1
  18     };
  19     if (line_number >= nrow(pm10_df)+1
  20         || strsplit(as.character(pm10_df[line_number,1])," ")[[1]][2] == '0:15')
  21     {
  22         break
  23     }}
  24     pm10_level = mean(pm10s, na.rm=TRUE)
  25     #Fill Pollution column: -1 if no info, 0 to 2 for pollution level
  26     if (!is.nan(pm10_level))
  27     {
  28         if (pm10_level < 30)
  29             meteo_df$Pollution[i] = 0
  30         else if (pm10_level <= 50)
  31             meteo_df$Pollution[i] = 1
  32         else #pm10 > 50
  33             meteo_df$Pollution[i] = 2
  34     }
  35
  36     #Also fill season + days of week variables
  37     meteo_df$Season[i] = ifelse(
  38         strsplit(as.character(meteo_df$Date[i]),'/')[[1]][1] %in% c("4","5","6","7","8"),
  39         1, 0)
  40     current_datetime = strptime(as.character(meteo_df$Date[i]), "%m/%d/%Y", tz="GMT")
  41     meteo_df$Week[i] = ifelse(current_datetime$wday %in% c(6,0), 0, 1)
  42 }
  43
  44 #see also:
  45 #http://stackoverflow.com/questions/8214303/conditional-replacement-of-values-in-a-data-frame
  46
  47 #Finally write new data
  48 write.csv(meteo_df, file="meteo_extra.csv", row.names=FALSE)