How to measure Kurtosis in Python pandas

Kurtosis! It's a neat statistical measure that tells you how different from a normal distribution a given set of data is. In particular it measures if data are heavy-tailed or light tailed when compared to a normal distribution.

The lower the number is, the less outliers exist in the data. The higher it is, the more outliers exist.

Let's take a look at the kurtosis for the price column in the following .csv of housing data.

homes_sorted.csv

How to measure kurtosis with Python pandas

Copy
import pandas as pd
df = pd.read_csv('/Users/kennethcassel/homes_sorted.csv')
df['price'].kurtosis()

Output: -0.29610470855022797

Conclusion:

It's super easy to analyze data to find kurtosis using python pandas.

Our dataset had a low kurtosis measurement. A normal distribution is 3. Anything below 1 is considered a light-tailed set of data. Anything higher than 1 is heavy tailed.