Data Architecture for Digital Banking

Abstract

This study aims to provide a data architecture for digital banking and to assess data quality. To reduce the risk that may arise from poor quality data in terms of money, reputation, and efficiency in using data for analysis and decision making and increase perspective on data usage to understand customer needs and present products or services precisely. Collect important and unique relevant data on the Data Lake system (January - August 2022) in 3 frequently used datasets. Then, it is used to develop the data architecture through the data management process using data quality management guidelines in accordance with the data governance policy of the Bank of Thailand (BOT).

The study found that there are three important data sets (Data Domain): customer data set, deposit data set, and loan data set. Important causes of data quality problems are 1) data is not up to date 2) data is not standardized and does not comply with data architecture principles and data quality standards, such as incorrect and incomplete data fields. In the case of a loan, if you are a natural person, you must have your name and ID card number. For legal entity customers, we will classify legal entities such as companies, department stores, stores, etc. Then process the data quality assessment (Data Quality Scorecard), collect data and report the results of the data quality assessment. To increase efficiency in controlling and tracking data in managing data quality problems for administrators and taking steps to notify data owners or relevant agencies, such as correcting data under the specified SLA.