A CLUSTER ANALYSIS AS THE BASIS FOR A PROFITABILITY ANALYSIS OF BIM PROJECTS

Building Information Modeling (BIM) has become increasingly popular in the German construction industry, particularly in transport infrastructure, to achieve project goals more reliably and increase productivity.Due to strong political demand to expand BIM applications, authorities need to demonstrate its added value, especially concerning cost-effectiveness.Existing approaches to examining BIM's economic efficiency often do not account for the specific conditions of German road construction and focus on individual projects, making it difficult to generalize results.As part of a research project commissioned by DEGES GmbH on behalf of the Federal Republic of Germany, an economic feasibility study on the application of BIM methodology in the planning and construction phases of federal trunk road construction is being conducted.With the approach of a project comparison of BIM and non-BIM projects, however, groups of similar projects are required.This article demonstrates how to establish the basis for this profitability analysis of BIM projects.Starting from a database of several hundred projects, the aim is to identify the most homogeneous project groups to serve as a basis for retrospective project analysis.To achieve this goal, a cluster analysis was conducted, focusing on determining which features are suitable for the technical characterization of projects and how these must be prepared.The cluster algorithms K-Means and DBSCAN were implemented and examined regarding optimal parameter settings.The cluster solutions were then interpreted, compared, and an algorithm was selected.The results show that, provided sufficient data is availableboth datasets and characteristics the use of cluster analysis is a useful tool for forming homogeneous groups of construction projects.