Get average by months of a time series (all Januaries, all Februaries, etc)

问题

I have a time series of daily data from 1992 to 2018. So far I have converted to monthly data but I also need to obtain anomalies per month and I need to obtain the average of each month over all years to finish with 12 averages. One for each month from each individual average of each year.

I have done the following using Pandas:

df = pd.read_excel(filename, "Daily", index_col=0)
df = df.resample("M").mean()

I have been trying to find how out to obtain now the average of each month every the whole time series but I have not found a way.

EDIT:

My data looks like this after resampling daily to monthly

    1         2         3   ...        37        38        39
1992-01-31  0.306511  0.310543  0.211181  ...  0.352130  0.348108  0.304041
1992-02-29  0.306236  0.312186  0.211741  ...  0.353696  0.343948  0.311114
1992-03-31  0.259254  0.297998  0.195577  ...  0.329181  0.294966  0.278523
1992-04-30  0.229502  0.297078  0.186802  ...  0.298462  0.267629  0.249950
1992-05-31  0.188347  0.240783  0.159703  ...  0.251465  0.215796  0.205284
1992-06-30  0.150345  0.213644  0.129967  ...  0.220702  0.179280  0.178172
1992-07-31  0.144945  0.213217  0.118467  ...  0.224497  0.163502  0.171851
1992-08-31  0.145402  0.188320  0.115089  ...  0.209280  0.159910  0.158608
1992-09-30  0.151685  0.194237  0.123106  ...  0.216324  0.174529  0.154490
1992-10-31  0.169207  0.235069  0.129761  ...  0.240324  0.197842  0.172253
1992-11-30  0.223199  0.271601  0.175349  ...  0.280514  0.258155  0.223209
1992-12-31  0.241892  0.302605  0.192563  ...  0.328505  0.289020  0.256858
1993-01-31  0.263852  0.351839  0.207057  ...  0.362024  0.340665  0.278063
1993-02-28  0.309779  0.392905  0.244505  ...  0.374407  0.386738  0.330977
1993-03-31  0.301839  0.364442  0.230318  ...  0.377743  0.344132  0.336906
1993-04-30  0.271325  0.317197  0.209343  ...  0.345088  0.306911  0.289592

(Date is the index, not column 1. Column 1 starts with 0.306511, and so on) All the way to end of 2018. So I need to obtain the average of all the Januaries, all the Februaries, etc. for each one of the columns.

回答1:

What you need is groupby:

m = df['Date'].dt.month
result = df.groupby(m).mean()

# Rename month 1 to January, 2 to February, etc.
result.index = pd.date_range('1/1/2019', '12/1/2019', freq='MS').strftime('%B')

Result (with random input):

               Value
January    51.838811
February   51.455804
March      51.275257
April      52.027894
May        49.101480
June       51.866638
July       51.600765
August     50.416463
September  48.732991
October    51.477874
November   50.797786
December   51.003006

来源：https://stackoverflow.com/questions/57914948/get-average-by-months-of-a-time-series-all-januaries-all-februaries-etc

标签

python

pandas

dataframe