Experiment proposal for data quality assessment in construction management

In recent years the usage of digital tools in construction management has increased, and with it the need for automation and applications of Artificial Intelligence (AI). The basis for AI to work properly is clean and meaningful data. However, several publications have appeared in recent years documenting the lack of good data in the construction industry. Most of the previous studies discussing the use of AI in construction management do not take into account the assessment or definition of data quality. Therefore, this paper presents an experiment to measure data quality of three construction management methods: Critical Path Method (CPM), Location Based Management System (LBMS) and Takt Planning and Takt Control (TPTC). Since location-based techniques (LB), such as LBMS and TPTC, require information on a higher level of detail for input than activity-based techniques—especially in terms of connecting processes, times, and locations—the expectation is that they also provide better data as output. This lead to the following hypothesis: construction projects using TPTC or LBMS provide a higher data quality for AI than projects using CPM. The proposed experiment consists of five phases: Project Definition, Project Planning, Tool Selection, Data Entry & Export, and Evaluation. The experiment was carried out to find LB methods do indeed provide a higher score across all defined evaluation metrics. However, the authors identified several limitations for the presented results, with the strongest one being a small sample size of evaluated software solutions, which leads to a strong bias towards the individual implementations. Therefore, the next stage of the authors’ research will be a more thorough execution of the experiment to verify the results.