Investigating profitability performance of construction projects using big data: A project analytics approach
Investigating profitability performance of construction projects using big data: A project analytics approach
复制标题
DOI：
10.1016/j.jobe.2019.100850
复制
发表时间：
2019-11-01
影响因子：
6.4
通讯作者：
Delgado, Juan Manuel Davila
中科院分区：
文献类型：
--
作者：
Bilal, Muhammad;Oyedele, Lukumon O.;Delgado, Juan Manuel Davila
建筑行业从项目开始阶段到项目交付阶段产生不同类型的数据。这些数据以各种形式和格式出现，超越了行业内使用的现有项目智能工具的数据管理、集成和分析能力。项目生命周期中的几项任务对建设项目的有效规划和交付具有影响。设定合适的利润率并随着项目进展持续跟踪利润率是至关重要的管理任务，需要数据驱动的决策支持。现有的利润估计方法使用公司或行业范围的基准来指导这些决策。这些基准通常是不可靠的，因为它们没有考虑到项目特定的变化。因此，使用统一费率对项目进行错误估计，最终由于支出不足或超支而导致完全不寻常的利润率。该研究提出了一种项目分析方法，利用大数据来了解不同类型建筑项目的盈利能力分布。为此，建议使用大数据架构，并展示了存储和分析大量项目数据的原型实现。我们的数据分析显示，利润率会不断变化，盈利能力表现在几个项目属性中各不相同。这些见解应作为知识纳入机器学习算法，以准确预测项目利润。所提出的方法使数据的快速探索，以了解不同类型的建设项目的盈利能力表现的基本模式。
The construction industry generates different types of data from the project inception stage to project delivery. This data comes in various forms and formats which surpass the data management, integration and analysis capabilities of existing project intelligence tools used within the industry. Several tasks in the project lifecycle bear implications for the efficient planning and delivery of construction projects. Setting up right profit margins and its continuous tracking as projects progress are vital management tasks that require data-driven decision support. Existing profit estimation measures use a company or industry wide benchmarks to guide these decisions. These benchmarks are oftentimes unreliable as they do not factor in project-specific variations. As a result, projects are wrongly estimated using uniform rates that eventually end up with entirely unusual margins either due to underspends or overruns. This study proposed a project analytics approach where Big Data is harnessed to understand the profitability distribution of different types of construction projects. To this end, Big Data architecture is recommended, and a prototype implementation is shown to store and analyse large amounts of projects data. Our data analysis revealed that profit margins evolve, and the profitability performance varies across several project attributes. These insights shall be incorporated as knowledge to machine learning algorithms to predict project margins accurately. The proposed approach enabled the fast exploration of data to understand the underlying pattern in the profitability performance for different types of construction projects.