0
$\begingroup$
  • I am trying to built a function to calculate the Geometric mean for Growth column and Avg_growth.
Year Company Name FH_PBIDT Growth Avg_growth
2009-10 Aayush Food & Herbs Ltd. 0.044881 0.000000 -0.773403
2010-11 Aayush Food & Herbs Ltd. 0.063545 0.415839 -0.773403
  • For that, I am using the below code:

from scipy.stats.mstats import gmean

  • I created a column name:

df['Geometric Mean'] = gmean(df.iloc[:,3])

df['Avg_Growth Mean'] = gmean(df.iloc[:,4])

  • I am getting an error that gmean is not defined. Could someone help me here?
$\endgroup$
2
  • $\begingroup$ What is your exact error? $\endgroup$ – HS-nebula Dec 8 '20 at 19:31
  • $\begingroup$ Error i am getting AttributeError: 'Series' object has no attribute 'gmean' $\endgroup$ – HARIKRISHNA MUDIDUDDI Dec 9 '20 at 4:26
1
$\begingroup$

Can also write a geometric mean function by a little help from Numpy.

import numpy as np

def geometric_mean(i):
    x = np.array(i)
    return x.prod()**(1.0/len(x))
$\endgroup$
0
$\begingroup$

As far as your error 'Series' object has no attribute 'gmean', make sure you are not doing this anywhere in your code:

# wrong way of calling gmean on a series
df.iloc[:,4].gmean() # accessing gmean as a function on series object

Instead how you have called gmean in your question description above is correct and works:

df['Avg_Growth Mean'] = gmean(df.iloc[:,4]) #works!

However, there might be more than one thing wrong here,

I tried testing code on data given by you and while gmean is available from the scipy.stats.mstats module, and it can be executed; the data is not apt for calculation of a geometric mean.

There is a zero value in Growth column, and negative values given in Avg Growth column. You can not find Geometric mean in case of negatives and zeroes: see this link for more details

See the image below for code I executed: Jupyternotebook screenshot with code

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.