JoAmps/Churn-prediction-in-a-vehicle-insurance-company-in-Ghana

Predicting customer churn in a vehicle insurance company in Ghana, deployed on Heroku, via CI/CD using GitHub Actions and set up monitoring to detect model and concept drifts # Project Name Churn prediction in a Vehicle Insurance Company in Ghana ## Project Intro/Objective Customer churn prediction is an important business activity for any company. The ability to predict this, is very key, so that companies identify such individuals and make the neccesary arrangements to keep them. Since this insurance company makes its profit from the number of customers that subscribes to its insurance plans, its absolutely crucial, they know which customers are likely to leave them, so they make the neccesary arrangement to keep them, in order not to lose money and keep making profits. ### Methods Used * Data exploration/descriptive statistics * Data processing/cleaning * Inferential Statistics * Machine Learning * Data Visualization * Predictive Modeling * Testing * Deployment * Monitoring ### Technologies * Python * Various python libraries for data science and machine learning * Heroku * Data Version Control(DVC) * Visual studio code, jupyter * Git * Testing(pytest) * CI/CD ## Project Description #### As mentioned in the Project objective, predicting these customers that are likely to churn is crucial, thats why this insurance company, needs to know the customers that would cancel their subscription to the insurance they give. #### Since this project is just a practice project to demonstrate my skills, i obtained this synthesized data from the insurance company, to use for my purposes. I obtained two data sets from them, The first was data about their customers from october 2021 to febrary 2022, and the second one was data about their customers in march 2022. ### The questions i deemed to explore were: #### What are the most crucial factors that affect the customer cancelling their subscription with the company. #### How strong is the relationship between the features(gender,type of plan,types of vehicles used etc), and whether they churned #### Which model could be beter suited to predicting this, and be can generalize well on new customers, a …