Optimizing construction project performance risk and contractor default prediction using resource management data

This article formulates a predictive risk scoring model of construction projects with logistic regression to predict contractor defaults and risks of project performance. In the analysis, a dataset provided by Kaggle is used with important variables of labor requirement, equipment usage, number of materials, duration of project, and efficiency of resource allocation. Preprocessing steps such as the management of missing data, data normalization, and encoding of categoric variables were done to get the data in the state to model. The logistic regression model was trained to estimate risks levels of the project with validation being done on the basis of accuracy, precision, recall, and F1-score. Among the most important findings, it can be noted that such aspects of the project as Mean Resource Demand and Material Quantities are relevant predictors of project risk. The model had a moderate predictive power with the accuracy of 36. The research paper is relevant to construction project management since it offers a numerical method of risk prediction and decision-making and recommends further investigation aimed at improving the accuracy of predictions with due to the use of more sophisticated models and larger data sets.