The rapid development of the technology industry has resulted in many telecommunication companies and internet service providers which can lead to competition. This can also make customers change providers so that it is interpreted as customer churn. Detecting customer churn from the beginning can make a company maintain its market.

In this project, we create a machine learning model to predict the customer churn. So, the model can help telecommunication companies and internet service providers to detect early signs of potential churn so that they can reduce customer losses and implement effective retention strategies.

The data shows that there are 3652 data is not churn and 598 data indicated churn.

After modeling, Random Forest Model has the highest accuration, but the difference between the number of false negatives and false positives is greater than Decission Tree. Thus, it can be concluded that the best model for predicting customer churn is the Decision Tree.

Meanwhile, the variable that influences the determination of customer is churn or not is total charge.