How does a major airport operator fold machine learning and AI into traditional business intelligence and actually get decisions out the other side? This session walks through how the Port of Portland does it, using two real examples: a passenger-arrival forecasting model for the PDX terminal, and a Customer Experience Scorecard. The talk follows the data end to end, from raw operational sources through Databricks pipelines and into Power BI reports that frontline business users rely on. Along the way it digs into the practical design choices that decide whether an AI-generated data product gets trusted or ignored: how to present model output so it reads as intuitive, how to convey uncertainty honestly, and how to make advanced analytics actionable rather than decorative. Hosted online by the Portland Fabric and Power BI User Group.
