Salon 2
Tell Me A Story: Using Advanced Statistical Techniques to Inform Practical Business Decisions (303231)
Aisha Mahmoud-Perez, Walmart Global People Analytics*Kalifa Oliver, Walmart Global People Analytics
Keywords: Random forest, structural equation modeling, logistic regression, principal components analysis, stress, turnover, satisfaction, inclusion, employee, attitude
Increasingly businesses are turning to “big data” to help inform complex business decisions. This practice is further complicated when companies want the data to be used to inform decisions that target factors affecting employees such as inclusion, satisfaction, stress, and turnover intentions. Given the implicit difficulties in trying to predict human behavior, as well as the limitations involved in collecting subjective data about employees, it is imperative that the statistical techniques used to analyze data, produce actionable and insightful findings, and subsequently inform business decisions be sound, defensible, and valid.
This presentation will look at four kinds of analysis used at a Fortune 100 company as we attempt to understand the drivers of our employees’ experiences at work. In modeling the data we used Principal Components Analysis, Random Forest, Logistic Regression, and Structural Equation Modeling. We will look at the strengths and weaknesses of each method and the utility of each technique at different stages of the item development and analysis process. The potential challenges faced when translating the technical aspects of data will also be discussed.