Scholarlink

<p>Given the importance of timely completion of roadway projects, numerous studies have investigated the underlying reasons for schedule overruns. However, the impacts of socioeconomic factors on schedule overruns are not known. Hence, this study utilized machine learning techniques including XGBoost, CatBoost, LightGBM, random forest, and logistic regression, applied to a dataset of 288 projects completed between 2018 and 2020 in Florida to uncover the impacts of socioeconomic factors on schedule overruns. CatBoost demonstrated the best performance, therefore the CatBoost classifier was used to develop a model that analyzes the impact of socioeconomic conditions, project characteristics, and other external factors on schedule overrun occurrences (SOO). The results were then visually represented using geographic information system (GIS). GIS integration was conducted using a 10-mile averaging radius for external factors, spatial joins and buffering in ArcGIS, and spatial correlation reduction, with projects visualized on a Florida map using color-coded overlays. The findings reveal the impacts of the key features on SOO, including annual average daily traffic (AADT) of trucks, original contract amounts, commuting-related factors, household-related factors, population of the wholesale industry, land use, and road direction. This study will assist planners in the transportation sector in anticipating the potential effects of socioeconomic factors on the occurrence of schedule overruns in planning and project control.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Investigating the impacts of socioeconomic conditions on schedule overrun occurrences in roadway projects: the fusion of machine learning and geospatial mapping

  • 期刊论文

  • 发表日期：

    2025年5月12日

查看全文
  • Alireza Shamshiri,
  • Kyeong Rok Ryu,
  • Mohsen Shahandashti,
  • June Young Park

摘要

Given the importance of timely completion of roadway projects, numerous studies have investigated the underlying reasons for schedule overruns. However, the impacts of socioeconomic factors on schedule overruns are not known. Hence, this study utilized machine learning techniques including XGBoost, CatBoost, LightGBM, random forest, and logistic regression, applied to a dataset of 288 projects completed between 2018 and 2020 in Florida to uncover the impacts of socioeconomic factors on schedule overruns. CatBoost demonstrated the best performance, therefore the CatBoost classifier was used to develop a model that analyzes the impact of socioeconomic conditions, project characteristics, and other external factors on schedule overrun occurrences (SOO). The results were then visually represented using geographic information system (GIS). GIS integration was conducted using a 10-mile averaging radius for external factors, spatial joins and buffering in ArcGIS, and spatial correlation reduction, with projects visualized on a Florida map using color-coded overlays. The findings reveal the impacts of the key features on SOO, including annual average daily traffic (AADT) of trucks, original contract amounts, commuting-related factors, household-related factors, population of the wholesale industry, land use, and road direction. This study will assist planners in the transportation sector in anticipating the potential effects of socioeconomic factors on the occurrence of schedule overruns in planning and project control.