Data Cleansing for Business Understanding
Information about business activities is usually captured in data describing on the one hand the structure of the business process (production view), on the other hand the activities of the users (customer view). Correct understanding of the performance of a business process depends essentially on the quality of the information in the different data and appropriate integration of various data sources. Frequently such activities are summarized under the heading Data Cleansing, which covers different types of procedures like data correction, record linkage or imputation of missing values. Due to the abundance of available data about a specific business process in traditional form as well as data on the Internet this is often a challenging task. In this lecture we present a unified process model for data cleansing and data provisioning and show how this model can be realized using the ADOxx platform. The basic idea of the model is simultaneous processing of the data workflow and the associated workflow of the metadata which describe the data processing activities. Such a model supports better understanding of the data and extends traditional methods for accessing data quality. As a use case for application we show how this approach can help in better monitoring of the COVID-19 pandemic.
Lecture at NEMO2022
Date/Time: Tuesday, July 12, 2022 at 11:30