Compute row average in pandas
Question:
Y1961 Y1962 Y1963 Y1964 Y1965 Region
0 82.567307 83.104757 83.183700 83.030338 82.831958 US
1 2.699372 2.610110 2.587919 2.696451 2.846247 US
2 14.131355 13.690028 13.599516 13.649176 13.649046 US
3 0.048589 0.046982 0.046583 0.046225 0.051750 US
4 0.553377 0.548123 0.582282 0.577811 0.620999 US
In the above dataframe, I would like to get average of each row. currently, I am doing this:
df.mean(axis=0)
However, this does away with the Region column as well. how can I compute mean and also retain Region column
Answers:
You can specify a new column. You also need to compute the mean along the rows, so use axis=1.
df['mean'] = df.mean(axis=1)
>>> df
Y1961 Y1962 Y1963 Y1964 Y1965 Region mean
0 82.567307 83.104757 83.183700 83.030338 82.831958 US 82.943612
1 2.699372 2.610110 2.587919 2.696451 2.846247 US 2.688020
2 14.131355 13.690028 13.599516 13.649176 13.649046 US 13.743824
3 0.048589 0.046982 0.046583 0.046225 0.051750 US 0.048026
4 0.553377 0.548123 0.582282 0.577811 0.620999 US 0.576518
If you are looking to average column wise. Try this,
df.drop('Region', axis=1).apply(lambda x: x.mean())
# it drops the Region column
df.drop('Region', axis=1,inplace=True)
I think this is what you are looking for:
df.drop('Region', axis=1).apply(lambda x: x.mean(), axis=1)
We can find the the mean of a row using the range function, i.e in your case, from the Y1961 column to the Y1965
df['mean'] = df.iloc[:, 0:4].mean(axis=1)
And if you want to select individual columns
df['mean'] = df.iloc[:, [0,1,2,3,4].mean(axis=1)
Taking the mean based on the column names
I am just sharing this which might be useful for those folks who want to take average of a few columns based on the their names, instead of counting the column index. This simply would be done using pandas’s loc instead of iloc. For instance, taking the odd-year average would be:
df["mean_odd_year"] = df.loc[:, ["Y1961","Y1963","Y1965"]].mean(axis = 1)
Y1961 Y1962 Y1963 Y1964 Y1965 Region
0 82.567307 83.104757 83.183700 83.030338 82.831958 US
1 2.699372 2.610110 2.587919 2.696451 2.846247 US
2 14.131355 13.690028 13.599516 13.649176 13.649046 US
3 0.048589 0.046982 0.046583 0.046225 0.051750 US
4 0.553377 0.548123 0.582282 0.577811 0.620999 US
In the above dataframe, I would like to get average of each row. currently, I am doing this:
df.mean(axis=0)
However, this does away with the Region column as well. how can I compute mean and also retain Region column
You can specify a new column. You also need to compute the mean along the rows, so use axis=1.
df['mean'] = df.mean(axis=1)
>>> df
Y1961 Y1962 Y1963 Y1964 Y1965 Region mean
0 82.567307 83.104757 83.183700 83.030338 82.831958 US 82.943612
1 2.699372 2.610110 2.587919 2.696451 2.846247 US 2.688020
2 14.131355 13.690028 13.599516 13.649176 13.649046 US 13.743824
3 0.048589 0.046982 0.046583 0.046225 0.051750 US 0.048026
4 0.553377 0.548123 0.582282 0.577811 0.620999 US 0.576518
If you are looking to average column wise. Try this,
df.drop('Region', axis=1).apply(lambda x: x.mean())
# it drops the Region column
df.drop('Region', axis=1,inplace=True)
I think this is what you are looking for:
df.drop('Region', axis=1).apply(lambda x: x.mean(), axis=1)
We can find the the mean of a row using the range function, i.e in your case, from the Y1961 column to the Y1965
df['mean'] = df.iloc[:, 0:4].mean(axis=1)
And if you want to select individual columns
df['mean'] = df.iloc[:, [0,1,2,3,4].mean(axis=1)
Taking the mean based on the column names
I am just sharing this which might be useful for those folks who want to take average of a few columns based on the their names, instead of counting the column index. This simply would be done using pandas’s loc instead of iloc. For instance, taking the odd-year average would be:
df["mean_odd_year"] = df.loc[:, ["Y1961","Y1963","Y1965"]].mean(axis = 1)