By: Eric Siegel, Program Chair, Predictive Analytics World for Business
In anticipation of his upcoming conference presentation at Predictive Analytics World for Business Las Vegas, June 16-20, 2019, we asked Allan Sammy, Director, Data Science and Audit Analytics at Canada Post, a few questions about incorporating predictive analytics into business. Catch a glimpse of his presentation, Auditing Analytic Models, and see what’s in store at the PAW Business conference in Las Vegas.
Q: You are a data scientist but you are an internal auditor. Can you explain that?
A: The role of Internal audit in any organization is to help management assess and manage risks that may prevent the organization from achieving its strategic and operational objectives. Today, one of the biggest risks facing organizations is a failure to embrace Analytics and maximize the value of data. Internal Audit can identify and quantify this Analytics Underutilization Risk, recommend the adoption of an analytics solution and in some cases even specify the appropriate type of model.
Q: How does Internal Audit work with the organization’s existing predictive analytics/models?
A: Predictive analytics is becoming more mainstream and therefore more subject to oversight and scrutiny. Management expects that model development be conducted in a structured, auditable fashion and that outputs are explainable and actually achieve the intended business objective. Internal audit applies a rigorous methodology to examining both the analytics development process and the inner workings of the model.
Q: How does predictive analytics deliver value at your organization – what is one specific way in which it actively drives decisions or operations?
A: As the largest delivery and logistics corporation in Canada, moving about 8.4 billion parcels and pieces of mail per year, we use predictive analytics in everything from volume forecasting to human resources and transportation optimization.
Q: What surprising discovery or insight have you unearthed in your data?
A: I can think of a few examples, however; in my current role as Director, Data Science and Audit Analytics and my previous role as Director, Fraud Risk Management, those examples typically involve unearthing the types of discoveries that can’t really be made public. For example, in my previous role my team and I built a neural network model to predict financial statement manipulation at service providers – and made several interesting discoveries.
Q: Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World.
A: For an analytics project to be successful in a corporate environment, developers have to be able to demonstrate to Management that their model has been created in a structured manner, achieves the intended objective and is able to withstand the audit process.
By: Eric Siegel, Conference Chair, Predictive Analytics World for Business