Application of business intelligence methods for analyzing a loan dataset

Abstract

The application of business intelligence methods is usually oriented on a specific type of business. This article is focused on the business intelligence data analysis of a sample loan dataset. There are a lot of well-known methods for analysing loan datasets. The research aim is to find some dependencies in a sample loan dataset which are not visible using techniques like sorting, filtering and pivoting. The following software products are used: Alyuda Neurointelligence and MS Power BI. As a result of the analysis it is proved that the amount of a loan depends mainly on the "born in the region" and "gender" factors. The analysis of the loan dataset in Alyuda Neurointelligence (using Input Importance) shows that the factor "born in the region" is the most important one affecting the amount of a loan. But the input importance shows only the possible strength of the influence of various factors on the amount of the loan and not the direction of their influence. The direction of their influence is found by applying the one-way ANOVA method in PSPP. The analysis of the loan dataset in Power BI shows some interesting dependencies. For debtors born outside the region, the difference in total loan amounts rises to over 4.5 times in favour of men. The average loan amount of debtors who were born outside the region is about 2.5 times greater than the average loan amount of debtors born in the region.

An Approach to Storing Data Based on the Data Lake Concept to Facilitate Intelligence Data Analysis

“…Their main components are: DW; ETL tools, online analytical processing (OLAP) techniques and data mining tools. BI methods are usually applied when analysing structured datasets from a specic type of business, e. g., banking or credit institutions [16,17]. Time series data are quite popular for storing data in dierent types of businesses.…”