Predicting Cost Overruns in Utility Streetworks
Léon olde Scholtenhuis
Organisation || BAM Energie en Water
Project Type || M.Sc. Thesis
Candidate || Rick Potkamp
Can cost overruns in utility projects be predicted based on enriched (historical) project data? This issue is becoming increasingly relevant as the number of projects in the utility sector is expected to rise due to, among other things, the energy transition. Therefore, it is desirable for a contractor to have a clear understanding of the risks of cost overruns. This research was conducted in three phases.
The first phase involves identifying influential features on cost overruns in utility projects based on literature and interviews with experts. This is done to determine which features affect such cost overruns. It may be that this project data is not yet being used or collected.
Stages in Project aimed at Predicting Cost Overruns in Streetworks
In the second phase, a database will be designed to combine project data and other datasets into one. After a feature importance analysis, it will be determined which data is relevant. As part of this, we will train a random forest and other regression models (e.g., LASSO).
Using these features, various machine learning methods will be applied, and models will be developed to optimally predict cost overruns. These will be evaluated, and finally, the usability of the model will be assessed by experts.