Get monthly means from dataframe of several years of daily temps

Question

I have daily temperature values for several years, 1949-2010. I would like to calculate monthly means. Here is an example of the data:

head(tmeasmax)
TIMESTE         


        

Answer 1

Here's a way to do it with the dplyr package:

library(dplyr)
library(lubridate)

tmeasmax$TIMESTEP = ymd(tmeasmax$TIMESTEP)

tmeasmax %>% 
  group_by(Year=year(TIMESTEP), Month=month(TIMESTEP)) %>%
  summarise(meanDailyMin=mean(MINIMUM.C.),
            meanDailyMean=mean(MEAN.C.))

  Year Month meanDailyMin meanDailyMean
1 1949     1       11.095      11.71928

You can summarise any other column by month in a similar way.

Answer 2

Here's a quick data.table solution. I assuming you want the means of MEAN.C. (?)

library(data.table)
setDT(tmeasmax)[, .(MontlyMeans = mean(MEAN.C.)), by = .(year(TIMESTEP), month(TIMESTEP))]
#    year month MontlyMeans
# 1: 1949     1    11.71928

You can also do this for all the columns at once if you want

tmeasmax[, lapply(.SD, mean), by = .(year(TIMESTEP), month(TIMESTEP))]
#    year month  MEAN.C. MINIMUM.C. MAXIMUM.C. VARIANCE.C.2. STD_DEV.C.      SUM COUNT
# 1: 1949     1 11.71928     11.095   12.64667     0.2942481   0.482513 1.426652     6

Answer 3

You can use the lubridate package to create a new factor variable consisting of the year-month combinations, then use aggregate.

library('lubridate')

tmeasmax2 <- within(tmeasmax, {
        monthlies <- paste(year(TIMESTEP),
                           month(TIMESTEP))
})

aggregate(tmeasmax2, list(monthlies), mean, na.rm = TRUE)