The goal of this project is to predict if an individual in given data are uninsured or not. The business goal is to reach out to the individuals who are predicted uninsured for marketing purposes.
We are given two datset: training_data.csv which has data of people and if they are insured or not. While we have unlabeled_data.csv which don't have any data on insurance. We are to predict uninsured or not on the later data.
The result of the project is to export a csv of person_id from the unlabeled data along with the predicted scores if the people are uninsured.
Once, the model is approved by my client, LetLife Insurance, we were asked to propose a strategy for how they should best engage with the audience of the model most likely to be uninsured. They’d like to understand how to better optimize their outreach across different outreach channels (eg. mail, ads, TV, internet, phone etc).