Data quality assessment for improved decision-making: A methodology for small and medium-sized enterprises

Industrial enterprises rely on prediction of market behavior, monitoring of performance measures, evaluation of production processes and other data analyses to support strategic and operational decisions. However, although an adequate data quality (DQ) is essential for any data analysis and several methodologies for DQ assessment exist, not all organizations consider DQ in decision-making processes. E.g., inaccurate and delayed data acquisition leads to imprecise master data and poor knowledge of machine utilization. While these aspects should influence production planning and control, current approaches to data evaluation are too complex to use them on a-day-to-day basis. In this paper, we propose a methodology that simplifies the execution of DQ evaluations and improves the understandability of its results. One of its main concerns is to make DQ assessment usable to small and medium-sized enterprises (SME). The approach takes selected, context related structured or semi-structured data as input and uses a set of generic test criteria applicable to different tasks and domains. It combines data and domain driven aspects and can be partly executed automated and without context specific domain knowledge. The results of the assessment can be summarized into quality dimensions and used for benchmarking. The methodology is validated using data from the enterprise resource planning (ERP) and manufacturing execution system (MES) of a sheet metal manufacturer covering a year of time. The particular application aims at calculating logistic key performance indicators. Based on these conditions, data requirements are defined and the available data is evaluated considering domain specific characteristics.