Applying the CRISP‐DM Framework for Teaching Business Analytics

Abstract

Experiential learning opportunities have been proven effective in teaching applied and complex subjects such as business analytics. Current business analytics pedagogy tends to focus heavily on the modeling phase with students often lacking a comprehensive understanding of the entire analytics process including dealing with real-life data that are not necessarily "clean" and/or small. Similarly, the emphasis on analytical rigor often comes at the expense of storytelling, which is among the most important aspects of business analytics. In this article, we demonstrate how the philosophy of the Cross Industry Standard Process for Data Mining (CRISP-DM) framework can be infused into the teaching of business analytics through a term-long project that simulates the realworld analytics process. The project focuses on problem formulation, data wrangling, modeling, performance evaluation, and storytelling, using real data and the programming language R for illustration. We also discuss the pedagogical theories and techniques involved in the application of the CRISP-DM framework. Finally, we document how the CRISP-DM framework has proved to be effective in helping students navigate through complex analytics issues by offering a structured approach to solving real-world problems.

Logistic Regression via Excel Spreadsheets: Mechanics, Model Selection, and Relative Predictor Importance

“…There are 500 additional observations in cells B1017:M1516, which are used as a holdout sample to evaluate predictive performance. As noted by Jaggia et al (2020), the importance of using a holdout sample should be stressed during the instruction process. To evaluate the predictors with respect to their explanatory importance, all possible subsets OLS regression is applied.…”

Performing social media analytics with Brandwatch for Classrooms: a platform review

“…The deployment phase addresses the need to effectively communicate and transform the SMA results into relevant business insights and data-driven business actions. The final phase label, deployment, was inspired and adopted from the final step in the Cross Industry Standard Process for Data Mining, abbreviated as CRISP-DM (Jaggia et al 2020 ).…”

“…The primary objective of the final learning phase is to interpret the analysis results, clearly communicate the results to a diverse audience, and translate the knowledge into actionable business recommendations (Yazici 2020 ; Jaggia et al 2020 ).…”

“…This final phase emphasizes the development of data storytelling skills. “Data storytelling refers to crafting and delivering compelling data-driven stories to decision makers for the purpose of converting insights into actions” (Jaggia et al 2020 ). This requires the development of complementary skills in data analysis, data visualization, and the development of a compelling narrative (Jaggia et al 2020 ).…”

“…“Data storytelling refers to crafting and delivering compelling data-driven stories to decision makers for the purpose of converting insights into actions” (Jaggia et al 2020 ). This requires the development of complementary skills in data analysis, data visualization, and the development of a compelling narrative (Jaggia et al 2020 ). The data visualization should incorporate contrast such as size, color, and shape to make it easier for the audience to interpret and consume the information.…”