Churn is when a customer stops doing business or ends a relationship with a company. It’s a common problem across a variety of industries, from telecommunications to cable TV to SaaS, and a company that can predict churn can take proactive action to retain valuable customers and get ahead of the competition. This course will provide you a roadmap to create your own customer churn models. You’ll learn how to explore and visualize your data, prepare it for modeling, make predictions using machine learning, and communicate important, actionable insights to stakeholders. By the end of the course, you’ll become comfortable using the pandas library for data analysis and the scikit-learn library for machine learning.
Exploratory Data AnalysisFree
Begin exploring the Telco Churn Dataset using pandas to compute summary statistics and Seaborn to create attractive visualizations.
Preprocessing for Churn Modeling
Having explored your data, it's now time to preprocess it and get it ready for machine learning. Learn the why, what, and how of preprocessing, including feature selection and feature engineering.
With your data preprocessed and ready for machine learning, it's time to predict churn! Learn how to build supervised learning machine models in Python using scikit-learn.Making Predictions50 xpPredicting whether a new customer will churn100 xpTraining another scikit-learn model100 xpEvaluating Model Performance50 xpCreating training and test sets100 xpCheck each sets length50 xpComputing accuracy100 xpModel Metrics50 xpConfusion matrix100 xpVarying training set size100 xpComputing precision and recall100 xpOther model metrics50 xpROC curve100 xpArea under the curve100 xpPrecision-recall curve50 xpF1 score100 xp
Learn how to improve the performance of your models using hyperparameter tuning and gain a better understanding of the drivers of customer churn that you can take back to the business.Tuning your model50 xpTuning the number of features100 xpTuning other hyperparameters100 xpRandomized search100 xpFeature importances50 xpVisualizing feature importances100 xpImproving the plot100 xpInterpreting feature importances50 xpAdding new features50 xpDoes model performance improve?100 xpComputing other metrics100 xpFinal thoughts50 xp
In the following tracksMarketing Analytics with Python
DatasetsTelco Churn Dataset
PrerequisitesData Manipulation with pandas
Mark PetersonSee More
Senior Data Scientist at Alliance Data
Mark is a senior data scientist who holds degrees in Predictive Analytics, Agriculture Economics, and Animal Science. He has worked on a variety of big data and machine learning projects across the US and Latin America including customer churn, part failures, smart cities, and NLP. He's interested in using AI to improve business processes and lives.