Process analytic approaches play a critical role in supporting the practice of business process management and continuous process improvement by leveraging process-related data to identify performance bottlenecks, extracting insights about reducing costs and optimizing the utilization of available resources. Process analytic techniques often have to contend with real-world settings where available logs are noisy or incomplete. In this paper we present an approach that permits process analytics techniques to deliver value in the face of noisy/incomplete event logs. Our approach leverages knowledge graphs to mitigate the effects of noise in event logs while supporting process analysts in understanding variability associated with event logs.