Better way to show Left Skewed Distributions

14 Aug 2020

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

df = pd.read_csv('train.csv')

plt.figure(figsize=(12,8))
sns.distplot(df["target"].values, bins=50, kde=False)
plt.xlabel('Test', fontsize=12)
plt.title("Test Histogram", fontsize=14)
plt.show()

png

This is a left skewed distribution. The basis of scoring will vary if this is what they give. Best way to deal with it is show it in a log scale

import numpy as np #For log function
plt.figure(figsize=(12,8)) #Define size of figure
sns.distplot(np.log1p(df["target"].values), bins=50, kde=False)
plt.xlabel('Test', fontsize=12)
plt.title("Log of the Test Histogram", fontsize=14)
plt.show()

png

Looks much better now