Assessment Task 1 – SAS Predictive Business Analytics

DATA EXPLORATION

1. Business Problem

With the increasing popularity of passive income strategies implemented to assist with long term secured savings, many banks have opportunities to launch direct marketing campaigns targeted at potential clientele, who prefer regularity in their returns, due to fixed interest rates. Generally the target customer for these campaigns is someone planning for long term returns, instead of short-term cash flow injection, as specified within a standard fixed term deposit.

Unique records within the ‘BANK_DIRECT_MARKETING’ dataset explicitly documents real-world examples of these direct marketing campaigns, and is a subset of a larger dataset extracted from a Portuguese bank, over a time period of two years, from May 2008 to November 2010 (Moro et al., 2011). The total number of unique records amounts to 10,578 entries [VERIFY]. During this time, the bank documented data relating to 17 marketing campaigns, targeted at convincing potential customers to enter a term deposit contract, via in-house telemarketing (Moro et al., 2011).

Using this dataset, our goal is to provide meaningful insights based on 18 input attributes such as job type, education, and marital status, mentioned in section 3.1 – data dictionary. Ultimately, upon completion of initial data exploration of attribute data, we aim to highlight key insights or trends within the data, and describe every attribute in respect to their distribution within the dataset. These steps will create the basis for further modelling and prediction steps within the data mining process; to eventually predict which customers would be more likely to agree to a term deposit, and help streamline future marketing campaigns.

2. Report Structure

This report aims to present key insights and observations derived from a comprehensive examination of the selected ‘BANK_DIRECT_MARKETING’ data set (UCI Machine Learning Repository, 2012), which will be documented below. The report is segregated into three major sections: data exploration, comparative analysis, and data pre-processing. Within the data exploration section, attributes will be individually categorised, highlighting their distribution, along with key metrics. Building upon this, comparative analysis will delve deeper into the data by providing bivariate and multivariate relationship insights for certain attributes. These comparisons will provide a basis for any required pre- processing that will occur in the following section. Lastly, a summary of findings will be collated to document useful trends identified in the first three sections of the report; as well as future objectives and challenges faced during exploratory data analysis. All observations catalogued below were obtained with the use of the SAS Viya for Learners cloud platform.