مدل داده کاوی برای شناسایی متغیرهای سود دهی پروژه

Many engineering design companies collect data such as profits to manage projects. But the relationships between operational variables and performance are usually not thoroughly analyzed and interpreted. This paper proposes a data mining model and procedure to relate influence variables to project profitability. Data categories and variables are defined at the project input, process and output stages. The model proposed herein was tested by analyzing 548 projects of an engineering consulting company. The relationships between profitability and various input and process variables were identified and interpreted. For example, the effect of QA/QC on profitability is positive. Based on documented data and derived information, this model can help companies gain operational knowledge and thus further improve performance.

Engineering companies usually collect data such as project profit and man-hours to manage design work but then do not analyze these data rigorously. Therefore, information is not derived to interpret cause–effect relationships, improve performance, plan future work, and create knowledge. Such inadequate engineering management is common and causes one-third of A/E projects to miss cost and schedule targets [1], [2] and [3]. Many studies propose frameworks to create or acquire knowledge, but they usually do not provide details of how this can be accomplished [4]. Knowledge management demands better information in terms of data needs, collection, analysis, and interpretation. When data needs and collection are well planned, analysis and interpretation will generate knowledge and better project management. For example, if project man-hour is seen as related to profitability, project man-hour and profit data are needed, collected, and analyzed to interpret their relationship to verify the tacit knowledge. For engineering companies, explicit knowledge is more technical in nature and can be more easily expressed and transmitted than tacit knowledge [5]. Tacit knowledge is difficult to articulate and mostly embedded in less-understood non-technical issues such as project management methods [6]. Bloodgood and Salisbury [7] argue that making such tacit knowledge available throughout the company will improve the company’s performance and profitability. Knowledge can be generated by data mining tools [8]. Data mining refers to the application of acquisition methods to the generation of potentially useful knowledge from the organization and analysis of raw data [9]. The research described in this paper establishes a system model and procedure to guide the data mining process. The model was tested by analyzing 548 projects of an engineering consulting company to search for the relationships and knowledge among the input variables, process variables, and profitability

Engineering project management needs to analyze cause–effect relationships more rigorously in order to get insight about operation and improve performance. This research proposes a data mining model and procedure to systematically analyze profitability variables at the project operation. Three categories: work data (I1) and work nature (I2), work division (P1) and management (P2), as well as performance (O) are established at a project’s input, process and output stages, respectively. A company has to define their own variables under these categories and collect sufficient data before using this model. The model was tested on 548 project dataset from an engineering design company. The variables of project data at the input stage and work division and management at the process stage are found to influence profitability. For example, (1) transportation projects are found more profitable than other types of projects; (2) construction supervision projects are more profitable than design and planning ones; (3) profitability has a negative relationship with project duration; (4) the effect of QA/QC on profitability is positive; and (5) project implementing QA/QC have lower U&E. It deserves the company’s further investigation to identify the reasons for them. The company did not analyze collected data before but became aware of the identified relationships after they were presented. The identified I–O, P–O, I–P, and I–P–O relationships provide insights about work operation mostly for the studied company. Other companies can find their own patterns by using this model and procedure. For easy understanding, only simple comparison and statistical methods are used in the analysis and explanation. Advanced data mining tools and software can be used to derive more insightful results based on this model and procedure. From the interpretation and inference made in the prior sections, the proposed model and procedure provides a vehicle to identify and further induce work knowledge. Explicit data and information such as contract amount and profitability are recorded frequently by a company, but mostly contained in documents that have not been transformed into knowledge. A company’s work process knowledge is usually stored in the brains of employees, not on paper or in a computer file. Guided by this systematic model and procedure, knowledge can be transformed from information through purposeful analysis and cause–effect relationship identification. Tacit knowledge can also be made explicit by soliciting opinions from experienced employees when interpreting the cause–effect relationships.