A data analytic approach to bank marketing - case study

Abstract

The data is related with direct marketing campaigns (phone calls) of a Portuguese banking institution. The classification goal is to predict if the client will subscribe a term deposit. Total 41,188 Clients record are available in the dataset. Key Data elements of the dataset are following – 1. Client Demographics (Age, Job, Marital & Education Status) 2. Client Financial Info (Default, Personal & Home Loan Status) 3. Campaign Attributes (Mode, Month/Day and Duration of contact) 4. Output variable / Desired target: Has the client subscribed a term deposit The objective of the analysis is to assess client details subscribing the Product – Bank Term Deposit and provide an insight to run an efficient & successful marketing campaign in future. We will analyze the average age of the clients subscribing the product and job type of the clients who subscribed the product. Then we will build a Machine Learning (Categorical) model which can predict whether the client will subscribe to a Term Deposit or not. As the dataset is highly imbalanced, so we applied smote and random oversampling methods and then we used logistic regression and xgboost and compared the accuracy, precision and recall. Finally, we saw how age and occupation can be useful in determining whether a client will subscribe a term deposit.