0

I have loop to analysis of 25 time series, In the analysis I need to get the mean of the individual columns for each station , for example the mean of all the Januaries, the same for all the Februaries, etc., etc. To complicate matters the time series are not all the same length, for example station 1 might run from 1900 to 1955 while station 2 might run from 1881 to 1945. So I need the mean of a Januaries of station 1, the mean of all Februaries of Station 1, etc., etc. and the same process for station 2 etc., etc. My times do not all start in January or end in December, but can start and finish in any month, each time series is individual. To get the colMeans, I need to change the time series to a matrix, but I need to pad out the empty spaces with NAs. Now how can I do that and put the function into the loop Below is an example of my data

       Jan   Feb   Mar   Apr   May   Jun   Jul   Aug   Sep   Oct   Nov   Dec
1951                                     15.50 18.74 22.75 25.90 25.43 27.61
1952 27.60 27.72 27.63 24.38 20.34 17.74 17.90 20.57 23.13 25.60 26.41 26.98
1953 25.80 26.19 24.99 23.23 19.59 15.78 14.85 18.97 20.44 25.78 26.65 27.00
1954 26.25 26.97 25.33 23.16 20.47 15.47 15.64 18.33 22.71 26.71 25.77 25.94
1955 26.69 25.36 24.19 23.42 19.65 17.10 17.36 18.67 20.95 24.41 24.93 26.12
1956 26.02 26.48 25.81 23.91 20.78 17.40 17.48 19.96 21.06 25.44 26.16 25.92
1957 26.67 28.03 25.24 24.40 19.89 16.54 17.99 19.01 24.81 26.18 28.38 26.96
1958 25.90 26.49 24.90 24.67 21.36 16.19 16.29 17.20 22.18 24.52 29.13 26.65
1959 26.53 26.65 25.17 24.26 20.67 17.56 18.11 18.49 21.50 26.21 26.48 27.52
1960 27.25 26.04 26.58 22.80 19.41 17.16 15.57 20.24 22.86 26.68 25.71 27.58
1961 26.79 25.88 26.19 24.22 22.09 17.77 17.91 18.56 23.27 24.94 25.68 26.66
1962 27.03 28.11 26.05 23.81 18.79 17.32 16.04 19.23 23.14 27.57 27.37 27.09
1963 26.91 26.68 24.97 22.87 18.71 16.79 16.05 18.25 23.52 25.73 27.08 26.86
1964 28.63 28.04 28.16 23.98 19.78 15.40 14.98 18.32 22.88 25.60 26.55 25.23
1965 27.77 28.87 26.62 23.40 19.49 15.62 17.14 19.79 22.09 23.44 26.32 28.40
1966 29.68 26.63 25.50 23.13 19.49 18.65 17.69 19.32 22.12 23.88 27.37 27.75
1967 27.84 26.46 25.75 24.20 20.15 17.22 15.64 18.39 22.41 25.38 27.42 27.62
1968 30.27 27.91 26.32 22.56 19.86 14.16 17.07 19.76 22.42 26.05 24.61 26.38
1969 28.72 30.04 25.85 23.68 20.09 18.32 16.85 19.61 22.10 24.97 27.28 25.46
1970 29.62 27.24 26.62 23.34 20.20 16.95 17.40 20.23 24.21 24.74 27.25 28.71
1971 26.25 26.44 28.15 25.31 19.14 16.21 16.92 19.12 23.09 24.28 24.43 27.19
1972 26.64 26.07 25.17 24.13 19.45 15.83 16.14 18.45 22.74 24.98 25.54 30.09
1973 30.44 28.54 28.10 22.80 20.36 18.05 16.74 19.16 23.56 24.64 25.97 25.50
1974 27.46 26.52 25.44 22.36 19.79 16.83 16.70 19.89 21.39 26.22 25.84 25.93
1975 27.84 25.91 24.42 22.87 20.90 16.42 16.25 18.79 23.24 24.54 27.07 26.97
1976 26.05 26.33 24.95 22.07 18.42 16.88 15.79 17.24 23.01 25.26 27.65 29.15
1977 30.00 26.31 23.47 23.17 20.21 17.23 16.12 18.80 23.55 25.93 26.97 26.68
1978 25.99 26.36 26.31 22.49 19.72 15.88 15.99 21.64 23.57 24.68 24.78 26.47
1979 26.98 28.17 25.36 24.06 20.65 17.64 15.98 19.89 22.63 25.40 25.31 27.03
1980 27.74 28.05 25.47 23.72 20.55 15.96
John Coleman
  • 51,337
  • 7
  • 54
  • 119
ian
  • 27
  • 5
  • What is the current form of the data? You want a matrix. What do you have now? – John Coleman Mar 04 '17 at 13:14
  • Currently the data is a ts time series.I want a way of getting the column means and then subtract the column from each observation of each column. A matrix wont work because I can't ensure that any or all ts will be multiple of 12 – ian Mar 04 '17 at 16:20
  • It seems that you have something like a list or vector of time series rather than a single time series. Please make a reproducible example, one that someone can copy into an R console and run. See http://stackoverflow.com/q/5963269/4996248 – John Coleman Mar 04 '17 at 16:24
  • It might help to reshape the data into a data frame with columns year, month, and value. If you did that, you could compute month summaries for all months concurrently. – Adam Bethke Mar 05 '17 at 13:57
  • Solved if guys with an if statement. But thank you for your ideas – ian Mar 07 '17 at 14:50

0 Answers0