How can I determine and remove outliers found by boxplot in python?

Question

I want to remove the outliers which are found by boxplot in my dataframe for each column. I know boxplot finds the outliers by IQR rule and displays them on graph. I know how to plot the boxplot using seaborn but I am unsure how can I determine exactly which rows these outliers actually refer to and how can I remove them ? Is there a function/method do to this ?

possible duplicate of https://stackoverflow.com/questions/23199796/detect-and-exclude-outliers-in-pandas-data-frame — Venkatesh Garnepudi, Jan 26 '19 at 07:41

score 3 · Answer 1 · edited Mar 08 '19 at 18:37

3

According to the basic definition of IQR outliers, Values less than Q1-1.5*IQR and values greater that Q3+1.5*IQR are treated as outliers. So,

Q1 =  df['col_name'].quantile(0.25)
Q3 = df['col_name'].quantile(0.75)
IQR = Q3 - Q1

Now, outliers are ,

df[(df['col_name'] < Q1-1.5*IQR ) | (df['col_name'] > Q3+1.5*IQR)]['col_name']

edited Mar 08 '19 at 18:37

squaleLis

6,116
2
22
30

answered Jan 26 '19 at 07:50

Venkatesh Garnepudi

316
1
10

How can I determine and remove outliers found by boxplot in python?

1 Answers1