Data drift refers to changes in data characteristics when comparing two distinct populations. In production systems, such shifts can impact downstream analytics and decision logic. In large-scale network environments, traffic patterns evolve continuously, and automated drift monitoring can complement traditional operational dashboards by focusing on distributional changes in the data.
This session uses network traffic data to demonstrate a practical approach for identifying and interpreting data drift. The presentation shows how key traffic attributes can be tracked over time and translated into clear, interpretable signals for analytics teams.
Attendees will gain insight into designing an automated monitoring workflow, interpreting drift signals in an operational context, and applying these insights to support more consistent and dependable analytics behavior over time.
Speaker(s): Anil Cavale,
Virtual: https://events.vtools.ieee.org/m/524412