Minimizing the data quality problem of information systems: A process-based method

Abstract

The low quality of data in information systems poses enormous risks to business operations and decision making. In this paper, a single-period resource allocation problem for controlling the information system's data quality problem is considered. We develop a Data-Quality-Petri net to capture the process through which data quality problem generates, propagates, and accumulates in the information system. The net considers not only the factors leading to the production of the data quality problem by the data operation nodes and the data flow structure, but also the data transfer ratio of the nodes. Then, we propose a nonlinear programming optimization model with control resource constraints. The result of the model provides an optimal strategy to allocate resources for minimizing the expected data quality problem of an information system. Further, we examine the impact of the data flow structure on optimal resource allocation. The result shows that the optimal resource input level for a data operation node is proportional to its potential for downstream propagation. A warehouse management system of an e-commerce company is utilized to illustrate the model. Our study provides a method for data managers to control the information system's data quality problem by employing a process perspective.