A Framework For Incorporating Data Mining Into An Accounting Curriculum

Abstract

As usage of data warehouses and data mining in corporate America increases, it is expected that accounting graduates will be called upon to contribute to data mining and knowledge discovery processes. Currently, the highly structured exercises and cases found in many accounting textbooks do not draw students’ attention to the complex task of mining large databases for information that might be used in business decision-making. This paper builds a typology of data mining tasks that considers the complexity of a database as well as the complexity of data mining tools, and proposes a multi-tier approach for teaching data mining concepts and techniques to accounting students. Drawing from that typology, the authors propose that accounting students be first introduced to simple databases that use flat files and simple data mining tools. As they advance through the accounting curriculum and learn about enterprise scale database management systems such as Oracle, accounting instructors can progressively introduce students to more complex data mining tasks. Eventually, these students can progress to complex data mining tasks that require denormalization of data tables and utilization of comprehensive, fully functional data mining tools. The paper suggests that instructors may use pivot table reports (e.g., Excel’s PivotTable reports) to provide accounting students with their first introduction to data mining. Product profitability analysis is used as an illustration. Although pivot tables would be at the low end in a typology of data mining tools, it is easy to prepare and use a pivot table to explore and summarize previously hidden patterns and relationships in a flat file. By using pivot tables, accounting students can summarize and analyze large data sets in a very familiar environment (e.g., Excel) and generate reports that allow them to apply filters and conveniently drill down to explore detailed patterns and relationships.