Customer Churn Prediction at Orange

EDA & prediction of customer churn at Orange with the use of Machine Learning. This project was done for practice in March 2020 using Google Colab.

The repository includes:

• colab_prediction.ipynb – Python notebook file extracted from Google Colab
• data
  • orange_score.csv – dataset of 10k rows to make the ML model
  • orange_train.csv – dataset of 100k rows to do the churn prediction
• results
  • list_of_all_customers_likely_to_churn.csv – list of all the 100k customers with the probability_of_churn score
  • list_of_the_100_customers_most_likely_to_churn.xlsx – the same list but limited to 100 rows in a descending order
  • pandas_profiling_report.html – Pandas Profiling report (as a quick/automated EDA)
  • presentation_to_the_business.pdf – 5-slide presentation about the findings for the Orange CEO
  • technical_report.pdf – 31-page technical report explaining the dataset, implemented approach, experimental setup, results and providing the final conclusions.