Collaborative Information Seeking for Construction Information Retrieval

During the construction process, constructors need to refer to large amounts of data and documents for making their daily decisions. These data and documents include job-site photos, construction schedule, on-site collected data about construction productivity, the history and plan of the crane operation. Most construction sites are dynamic and spatially complicated, hence these data tend to generate and be updated frequently, so that it might be challenging for constructors to seek useful information from these data to support their decisions, such as which parts of the constructed facility should be inspected in greater detail.

Due to the fragmental nature of the construction industry, the data and documents are scattered all places, and it is very difficult to organize them in a consistent manner despite the fact that many document organization development companies tried to do so (:::ref). Without a good organization, and since it is almost impossible to get such a uniform organization, constructors always rely on some search mechanisms to retrieve information from documents. Text-document search and image search algorithms are used by them to find useful information from large amounts of data and documents, that somehow address the problem of information retrieval for decision support in the building construction domain.

However, such ad-hoc approaches for retrieving construction related information might be subject to the search skills of the constructors, and it is likely that some constructors might not be able to extract all critical information related to their decisions. That might lead to construction accidents and late discovery of inconsistency of information, which can cause delays or failures of construction projects.

Some observation in some preliminary studies (::: references) indicate that rather than viewing information seeking as a single-user and single query sessions, there might be a potential to link these queries together, store these queries, and analyze these queries for future reference. This observation makes us believe that construction related information seeking do not have to be a single-person and single query process, and it is possible to enable collaborative search for more systematic, comprehensive, and efficient information seeking.

Recent progresses in the domain of collaborative search indicate that social network can be used to automatically suggest what queries might be useful for a specific user under a specific scenario. It has also been shown that the search history could be organized to tell people that under similar conditions what are the most critical information retrieved by previous people (including themselves). With such storage of correlations between queries of multiple people and multiple query sessions, it might be possible to systematically organize the information seeking "projects" for enabling more comprehensive information seeking for construction tasks and reducing problems caused by incomplete information seeking, and inefficient information retrieval.

Major advantages of such a collaborative and historic recording approach of construction information seeking could be: 1) systematic information organization, which might be more complete compared to single-query session approach; 2) the retrieved information can be labeled by people so that later they do not have to retrieve the same information again, which might be able to improve the information retrieval efficiency by saving the time for repeated searching sessions.