Search

The advancement of natural language processing and text mining techniques facilitate automatic non-trivial pattern extraction and knowledge discovery from text data. However, text-based research has received less attention compared to image- and sensor-based research in the construction industry. Hence, this paper performs a comprehensive review to understand the current state and future insights of text analytics focusing on the data source and analysis method. This study identifies various kinds of text data sources from project documents as well as open data in the websites. In addition, the review finds that the ontology- and rule-based approach has been dominant, at the same time, recent research has attempted to apply the state-of-the-art machine learning methods. It is envisioned that there are potential advancements in construction engineering and management based on the latest text analysis methods along with the enriched data by the digital transformation.

With the rising adoption of BIM in the AECO (Architecture, Engineering, Construction and Owner-operator) industry, BIM content is increasingly more complete and valuable. Unfortunately, the data is largely locked inside the respective BIM authoring tool with limited facility for users to perform queries beyond simple queries on objects and their properties. Even with the use of open data exchange format such as IFC (Industry Foundation Classes), getting the data out of BIM is still not an easy task. This research focuses on defining a schema and transformation rules to bring BIM data into a form that is easily queried. The schema is defined using popular relational database structure following an established star schema model that is well-defined within the data warehouse domain. The primary aim of the concept is to allow flexible and efficient queries into the BIM data using standard SQL that removes restrictions on predefined queries, effectively transforming BIM data into an open and queryable database. Another major contribution of this research is the significant potential for supporting spatial queries. This is a critical feature often neglected in the current approach to BIM-based databases. This paper presents the integration of spatial operations into standardized SQL queries making the BIM data accessible for wide ranges of query capabilities. Such feature will allow much better visibility into BIM data for better decision-making processes.