Tobias Hildebrandt

Something Doesn't Sound Right: Sonification for Monitoring Business Processes in Manufacturing

In manufacturing processes a problem during production, such as a e.g. A broken machine, can lead to a standstill and thus to a loss in revenue or even to a contractual penalty. Monitoring staff in manufacturing and in other industries therefore usually observe their business processes with systems that present process-related events and data by using different types of data visualization. This has several drawbacks, e.g. That either users cannot efficiently perform other tasks while observing their monitoring application, or in case they look at their screens only infrequently, that they risk to miss potentially time-critical events or alerts. Therefore we propose to combine current visual-based process monitoring systems with techniques from the area of sonification (the presentation of data using sound). Many factories already contain auditory alerts and alarms, but these usually do not convey a lot of information and are often considered to be obtrusive and distracting. We developed a sonification framework that can receive events from execution engines and preprocess and sonify them according to user-defined settings and filters. Different sonification prototypes for both, event-based and quantitative data (such as KPIs), have been developed and discussed during a focus group meeting with users and scientists from the domain of industrial management and production monitoring.

Testing Processes with Service Invocation: Advanced Logging in CPEE

Business process analysis is one of the major concerns for companies: before processes are enacted and executed, flaws and bottlenecks should be removed. Process simulation has been traditionally used to simulate process paths based on stochastic information. However, the obtained results are often not consistent with the behavior that can be observed later during process execution as the functioning and reaction of the orchestrated web services cannot be considered. To provide more realistic results, a process testing environment is presented that equips the Cloud Process Execution Engine (CPEE) with a newly developed logging component. The contribution of the logging component is that it incorporates simulated data from invoked services and thus allows for the generation of more realistic log files.

Exploration of the Potential of Process Mining for Intrusion Detection in Smart Metering.

Process mining is a set of data mining techniques that learn and analyze processes based on event logs. While process mining has recently been proposed for intrusion detection in business processes, it has never been applied to smart metering processes. The goal of this paper is to explore the potential of process mining for the detection of intrusions into smart metering systems. As a case study the remote shutdown process has been modeled and a threat analysis was conducted leading to an extensive attack tree. It is shown that currently proposed process mining techniques based on conformance checking do not suffice to find all attacks of the attack tree; an inclusion of additional perspectives is necessary. Consequences for the design of a realistic testing environment based on simulations are discussed.

The First Data Science Challenge at BTW 2017

The 17th Conference on Database Systems for Business, Technology, and Web (BTW2017) of the German Informatics Society (GI) took place in March 2017 at the University of Stuttgart in Germany. A Data Science Challenge was organized for the first time at a BTW conference by the University of Stuttgart and Sponsor IBM. We challenged the participants to solve a data analysis task within one month and present their results at the BTW. In this article, we give an overview of the organizational process surrounding the Challenge, and introduce the task that the participants had to solve. In the subsequent sections, the final four competitor groups describe their approaches and results.

Combining Sonification and Visualization for the Analysis of Process Execution Data

Business process execution data is analyzed for different reasons such as process discovery, performance analysis, or anomaly detection. However, visualizations might suffer from a number of limitations. Sonification (the presentation of data using sound) has been proven to successfully enhance visualization in many domains. Although there exist approaches that apply sonification for real-time monitoring of process executions, so far this technique has not been applied to analyze process execution data ex post. We therefore propose a multi-modal system, combining visualization and sonification, for this purpose. The concepts are evaluated by a prototypical ProM plugin as well as based on a use case.

Short Paper: Towards Enhancing Business Process Monitoring with Sonification

State-of-the-art business process monitoring systems usually base on different types of real-time visualizations, in which data is typ- ically presented using various graphical elements such as speedometers. However, these systems have several drawbacks, such as the inability to constantly monitor process executions while at the same time work- ing on other things. This is why this paper proposes to enhance visual process monitoring with techniques from the area of sonification (the presentation of data using sound). Even though sonification has already successfully been evaluated in several domains for real-time monitoring, there is so far no comprehensive research for its usage in business process monitoring. This paper proposes sonification techniques and user inter- actions that can be implemented in future applications.