mean<\/span>(<\/span>axis=<\/span>None<\/span>, skipna=<\/span>None<\/span>, level=<\/span>None<\/span>, numeric_only=<\/span>None<\/span>, **kwargs<\/span>)\r\n<\/span><\/pre>\nParameters::<\/p>\n
\n- Axis is the parameter on which the function will be applied. It denotes a boolean value for rows and column.<\/li>\n
- Skipna excludes the null values when computing the results.<\/li>\n
- If the axis is a MultiIndex (hierarchical), count along with a particular level, collapsing into a Series.<\/li>\n
- Numeric_only will use the numeric values when None is there.<\/li>\n
- **kwargs: Additional keyword arguments to be passed to the function.<\/li>\n<\/ul>\n
This function returns the mean of the values.<\/p>\n
Let’s dig in deeper to get a thorough understanding!<\/em><\/p>\nPandas: Replace NaN with column mean<\/h3>\n
We can replace the NaN values in the whole dataset or just in a column by getting the mean values of the column.<\/p>\n
For instance, we will take a dataset that has the information about 4 students S1 to S4 with marks in different subjects.<\/p>\n
![\"Pandas:](\"https:\/\/python-programs.com\/wp-content\/uploads\/2021\/04\/Screenshot-412.png\")
<\/p>\n
Code:\r\n\r\n<\/pre>\nimport numpy as np<\/span><\/div>\nimport pandas as pd<\/span><\/div>\n# A dictionary with list as values<\/span><\/div>\nsample_dict = <\/span>{<\/span> ‘S1’<\/span>: <\/span>[<\/span>10<\/span>, <\/span>20<\/span>, np.<\/span>NaN<\/span>, np.<\/span>NaN<\/span>]<\/span>,<\/span><\/div>\n‘S2’<\/span>: <\/span>[<\/span>5<\/span>, np.<\/span>NaN<\/span>, np.<\/span>NaN<\/span>, <\/span>29<\/span>]<\/span>,<\/span><\/div>\n‘S3’<\/span>: <\/span>[<\/span>15<\/span>, np.<\/span>NaN<\/span>, np.<\/span>NaN<\/span>, <\/span>11<\/span>]<\/span>,<\/span><\/div>\n‘S4’<\/span>: <\/span>[<\/span>21<\/span>, <\/span>22<\/span>, <\/span>23<\/span>, <\/span>25<\/span>]<\/span>,<\/span><\/div>\n‘Subjects’<\/span>: <\/span>[<\/span>‘Maths’<\/span>, <\/span>‘Finance’<\/span>, <\/span>‘History’<\/span>, <\/span>‘Geography’<\/span>]}<\/span><\/div>\n# Create a DataFrame from dictionary<\/span><\/div>\ndf = pd.<\/span>DataFrame<\/span>(<\/span>sample_dict<\/span>)<\/span><\/div>\n# Set column ‘Subjects’ as Index of DataFrame<\/span><\/div>\ndf = df.<\/span>set_index<\/span>(<\/span>‘Subjects’<\/span>)<\/span><\/div>\nprint<\/span>(<\/span>df<\/span>)<\/span><\/div>\nSuppose we have to calculate the mean value of S2 columns, then we will see that a single value of float type is returned.<\/p>\n
![\"Mean](\"https:\/\/python-programs.com\/wp-content\/uploads\/2021\/04\/Screenshot-414.png\")
<\/p>\n
Code:\r\n\r\n<\/pre>\nmean_value=df<\/span>[<\/span>‘S2’<\/span>]<\/span>.<\/span>mean<\/span>()<\/span><\/div>\nprint<\/span>(<\/span>‘Mean of values in column S2:’<\/span>)<\/span><\/div>\nprint<\/span>(<\/span>mean_value<\/span>)<\/span><\/div>\nReplace NaN values in a column with mean of column values<\/h3>\n
Let’s see how to replace the NaN values in column S2 with the mean of column values.<\/p>\n
![\"Replace](\"https:\/\/python-programs.com\/wp-content\/uploads\/2021\/04\/Screenshot-416.png\")
<\/p>\n
Code:<\/p>\n
\ndf<\/span>[<\/span>'S2'<\/span>]<\/span>.<\/span>fillna<\/span>(<\/span>value=df<\/span>[<\/span>'S2'<\/span>]<\/span>.<\/span>mean<\/span>()<\/span>, inplace=<\/span>True<\/span>)\r\n<\/span>print(<\/span>'Updated Dataframe:'<\/span>)\r\n<\/span>print(<\/span>df<\/span>)<\/span><\/pre>\n<\/div>\nWe can see that the mean() method is called by the S2 column, therefore the value argument had the mean of column values. So the NaN values are replaced with the mean values.<\/p>\n
Replace all NaN values in a Dataframe with mean of column values<\/h3>\n
Now, we will see how to replace all the NaN values in a data frame with the mean of S2 columns values.<\/p>\n
We can simply apply the fillna() function with the entire data frame instead of a particular column.<\/p>\n
![\"Replace](\"https:\/\/python-programs.com\/wp-content\/uploads\/2021\/04\/Screenshot-418.png\")
<\/p>\n
Code:<\/p>\n
\ndf.<\/span>fillna<\/span>(<\/span>value=df<\/span>[<\/span>'S2'<\/span>]<\/span>.<\/span>mean<\/span>()<\/span>, inplace=<\/span>True<\/span>)\r\n<\/span>print(<\/span>'Updated Dataframe:'<\/span>)\r\n<\/span>print(<\/span>df<\/span>)<\/span><\/pre>\n<\/div>\nWe can see that all the values got replaced with the mean value of the S2 column. The inplace = True has been assigned to make the permanent change.<\/p>\n
Pandas: Replace NANs with mean of multiple columns<\/h3>\n
We will reinitialize our data frame with NaN values.<\/p>\n
![\"Pandas:](\"https:\/\/python-programs.com\/wp-content\/uploads\/2021\/04\/Screenshot-420.png\")
<\/p>\n
Code:<\/p>\n
\ndf = pd.<\/span>DataFrame<\/span>(<\/span>sample_dict<\/span>)\r\n<\/span># Set column 'Subjects' as Index of DataFrame\r\ndf = df.set_index<\/span>(<\/span>'Subjects'<\/span>)\r\n<\/span># Dataframe with NaNs\r\nprint(<\/span>df<\/span>)<\/span><\/pre>\n<\/div>\nIf we want to make changes to multiple columns then we will mention multiple columns while calling the mean() functions.<\/p>\n
![\"Mean](\"https:\/\/python-programs.com\/wp-content\/uploads\/2021\/04\/Screenshot-422.png\")
<\/p>\n
Code:<\/p>\n
\nmean_values=df<\/span>[[<\/span>'S2'<\/span>,<\/span>'S3'<\/span>]]<\/span>.<\/span>mean<\/span>()\r\n<\/span>print(<\/span>mean_values<\/span>)<\/span><\/pre>\n<\/div>\nIt returned the calculated mean of two columns that are S2 and the S3.<\/p>\n
Now, we will replace the NaN values in columns S2 and S3 with the mean values of these columns.<\/p>\n
![\"replace](\"https:\/\/python-programs.com\/wp-content\/uploads\/2021\/04\/Screenshot-424.png\")
<\/p>\n
Code:<\/p>\n
\ndf<\/span>[[<\/span>'S2'<\/span>,<\/span>'S3'<\/span>]]<\/span> = df<\/span>[[<\/span>'S2'<\/span>,<\/span>'S3'<\/span>]]<\/span>.<\/span>fillna<\/span>(<\/span>value=df<\/span>[[<\/span>'S2'<\/span>,<\/span>'S3'<\/span>]]<\/span>.<\/span>mean<\/span>())\r\n<\/span>print(<\/span>'Updated Dataframe:'<\/span>)\r\n<\/span>print(<\/span>df<\/span>)<\/span><\/pre>\n<\/div>\nPandas: Replace NANs with row mean<\/h3>\n
We can apply the same method as we have done above with the row. Previously, we replaced the NaN values with the mean of the columns but here we will replace the NaN values in the row by calculating the mean of the row.<\/p>\n
For this, we need to use .loc(\u2018index name\u2019) to access a row and then use fillna() and mean() methods.<\/p>\n
![\"Pandas:](\"https:\/\/python-programs.com\/wp-content\/uploads\/2021\/04\/Screenshot-426.png\")
<\/p>\n
Code:<\/p>\n
\ndf.<\/span>loc<\/span>[<\/span>'History'<\/span>]<\/span> = df.<\/span>loc<\/span>[<\/span>'History'<\/span>]<\/span>.<\/span>fillna<\/span>(<\/span>value=df.<\/span>loc<\/span>