2021 EDSIG Proceedings: Abstract Presentation

Debriefing an SQL Exploratory Activity to Detect Data Anomalies, Summary Statistics, and Potential Fraud Within a 100,000-Invoice Dataset

Reagan Siggard
Utah State University

Robert Mills
Utah State University

Pamela Dupin-Bryant
Utah State University

David Olsen
Dixie State University

Abstract
According to Harvard Business Publishing editors, featuring Azeem Azhar and Geraint Rees, “we’re not teaching people to pass an exam; we’re teaching them how to think and how to deeply engage with the world around them” (2021). This abstract presentation, introduces a constructivist data exploration module using the 5E Instructional model. This module is intended to complement an existing database course based primarily on objectivist principles. The SQL-Explore Learning Module detailed in the presentation provides students with an opportunity to go beyond simply passing exams. Students are encouraged to explore the data to better understand the story it tells, including anomalies, patterns, trends, and potential fraud. Additional detailed benefits of this module are included in Table 1. --Insert Table 1 Here-- With this exploratory activity, three general investigative categories were identified as students worked to understand the large invoice dataset given in the SQL-Explore Learning Module. First, a straightforward approach was utilized by students who worked to respond to specific conditions given in the SQL-Explore Learning activity. Second, students who identified a trend in the dataset and worked to discover the origin of the trend. Lastly, students who created their own hypothetical company standards and found occurrences when those standards were violated. The purpose of this abstract presentation is to provide an overview of the constructivist exploratory activity and example SQL code from three general investigative categories utilized by students. The presentation will also provide details on how to integrate this module into existing database course curriculum, including the code to generate custom 100,000-row invoice datasets for student exploration. Finally, the authors will present initial student reactions and perceived learning based on the exploratory activity. The presentation will conclude with suggestions for improvements and future constructivist activities for database and other data analytics and information system courses.