Machine Learning Made Easy by PyCaret

Entire machine learning pipeline with 10 lines of code.

PyCaret is a python open source low-code machine learning library created by Moez Ali and released in April 2020. It is literally a low-code library which allows to create an entire machine learning pipeline with very few lines of code. PyCaret is essentially a wrapper built on common python machine learning libraries such as scikit-learn, XGBOOST and many more.

What PyCaret achieves is a higly simple yet functional syntax. For instance, we can compare 18 classification models with 1 line of code. In this post, I will walk you through a classification task using PyCaret and explain the details of each step.

Let’s start with installing PyCaret:

!pip install pycaret

If you use google colab as your IDE and plan to render interactive visualizations in the notebook, following code needs to be executed:

from pycaret.utils import enable_colab
enable_colab()

The dataset we will use is “ Telco Customer Churn ” dataset which is available on kaggle. After importing numpy and pandas, we can read the dataset into a pandas dataframe:

import numpy as np
import pandas as pddf = pd.read_csv("/content/Customer-churn.csv")
df.shape
(7043, 21)

The dataset has 7043 observations (rows) and 21 columns. Here is the list of columns:

“CustomerID” does not have any informative power since it is just a random rumber assigned to each customer. “TotalCharges” column is multiplication of “tenure” and “MonthlyCharges” columns so we don’t need this column as well. We just drop these two columns:

df.drop(['customerID','TotalCharges'], axis=1, inplace=True)