The Event Tunnel: Interactive Visualization of Complex Event Streams for Business Process Pattern Analysis

Abstract

Event-based systems are gaining increasing popularity for building loosely coupled and distributed systems. Since business processes are becoming more interconnected and event-driven, event-based systems fit well for supporting and monitoring business processes. In this paper, we present an event-based business intelligence tool, the Event Tunnel framework. It provides an interactive visualization of event streams to support business analysts in exploring business incidents. The visualization is based on the metaphor of considering the event stream as a cylindrical tunnel, which is presented to the user from multiple perspectives. The information of single events laid out in the Event Tunnel is encoded in event glyphs that allow for a selective mapping of event attributes to colors, size and position. Different policies for the placement of the events in the tunnel as well as a clustering mechanism generate various views on historical event data. The Event Tunnel is able to display the relationships between events. This facilitates users to discover root causes and causal dependencies of event patterns. Our framework couples the event-tunnel visualization with query tools that allow users to search for relevant events within a data repository. Using query, filler and highlighting operations the analyst can navigate through the Event Tunnel until the required information or event patterns become visible. We demonstrate our approach with use cases from the fraud management and logistics domain.