完整熊猫数据框的标准偏差和平均值

Question

I've got quite a large dataset and would like to calculate the mean and the standard deviation, across all columns and rows. Unfortunately, I haven't got a proper solution for this yet. My dataset looks a littlebit like that (total of 600 rows):

When I use the pandas-function weekl_rtr.mean() I just calculate the mean across each column. The workaround, that might work for the mean should be weekl_rtr.mean().mean() , but this does not work for the standard deviation. Do you have an idea, how to solve this?

Thank you and kind regards,

Markus

Answer 1

To my knowledge, there is no direct way to do it in pandas. You have two options:

1. Get the underlying numpy array and calculate mean or std on it. In contrast to pandas this will evaluate the function across all dimentions by default. For example, you can do df.values.mean() or df.to_numpy().mean() in pandas 0.24+.
2. Transform the table into a single column and then run the desired operation on that column

Answer 2

Change the axis for mean and standard deviation:

# Across columns (Default)
weekl_rtr.mean(axis = 0)
# or
weekl_rtr.mean()

# Across rows
weekl_rtr.mean(axis = 1)

The same applies to std() . You should also look into df.describe() which describes a DataFrame with more statistics info (mean, std, count, min, max, percentiles):

# Across columns
weekl_rtr.describe()

# Across rows
weekl_rtr.apply(pd.DataFrame.describe, axis=1)