Creating event logs from heterogeneous, unstructured business data

Abstract

Efficient processes give companies the edge required to prevail in global competition. Processes can have a high impact on important factors like product and service quality as well as overall economic efficiency. Hence process improvement plays an increasingly important role in many companies. The first step in process improvement and analysis is understanding the process. While a number of process analysis tools are available, these tools can only analyze processes for which log data (e.g. generated by BPM systems) exists. This paper introduces a tool that allows users to collect and structure traces from undocumented processes like workarounds or improvised processes in order to generate log files. The tool supports query specific ad-hoc exchange of ontologies in order to extract information from unstructured documents containing process traces as well as data extraction components for common databases. It thus bridges the gap between process traces in unstructured, heterogenous documents and process analysis software.