Using Basic Natural Language Processing for Effective Project Closeout Process: Fid-Bau Portal

Using Basic Natural Language Processing for Effective Project Closeout Process

Abdelaty, Ahmed / Nesselhauf, Kevin

Project closeout process is a crucial stage to complete every contractual obligation between the project stakeholders. However, the closeout process usually result in non-adding value activities to satisfy the contract’s requirement. In many cases of the project closeout process, key stakeholders work in a negative environment to finish the project’s requirements by generating and executing a list of punch list items. Many contractors consider the execution of punch list items including rework or missing work as non-value added activities. This perspective can create a strained relationship between the project’s stakeholders including subcontractors and clients. Therefore, this study aims at improving the project closeout phase by analyzing textual data of punch list item recorded by one construction company. The text analysis aims at determining the most recurring punch list items across different projects. Based on the results, a proactive approach to manage punch list items during the project closeout process is proposed to effectively minimize the number of punch list items for a given project. One case study is conducted to test the validity of the proposed approach. It is expected that the proposed approach will help contractors to formally document the project closeout phase as well as predict and proactively address defects in a timely manner.