Advanced eye tracking analysis for investigating construction craft professional interactions with 2D drawings

The workforce of skilled construction craftworkers in the United States is both shrinking, and aging, largely due to a lack of interest from young people to enter the industry. These trends have negatively impacted construction project productivity for several decades, and now, more than ever, innovative methods are needed to improve construction productivity. This research posits that productivity improvements could be realized through incremental improvements to traditional 2D construction drawings, the industry standard engineering information deliverable that has seen very little change in well over a century. Given the widespread use and familiarity of 2D construction drawings, any improvements to the format or to the use of the format could have immediate, major impacts on industry-level productivity. To that end, this research used eye tracking technology to investigate how professional pipefitters interacted with 2D drawings during a pipe model assembly task. Existing glance analysis methods were applied and expanded upon to develop an open-source eye tracking analysis web application, and the application was used to identify relationships between pipefitter demographics, experience, abilities, and performance. Initial efforts revealed that existing eye tracking glance analysis standards can lead to inconsistent results, so recommendations are provided for the further standardization of eye tracking glance/visit analyses. A new potential noise filtering parameter was introduced, maximum off-stimulus fixations, which could be incorporated into future studies. Convex hull coverages were also used to measure the efficiency in which the pipefitters were able to search for information within the drawings, and the complexity of visual information in the drawings was also quantified. Significant linear relationships were observed between information complexity and search efficiency, but the most noteworthy findings were the consistent relationships observed between the metrics of search efficiency and performance. The results suggest that future efforts could improve craftworker productivity by modifying 2D drawings or by training craftworkers to read 2D drawings in ways that improve their search efficiency. The eye tracking analysis application developed for this work has been made freely available, and this work concludes with many suggestions for future studies.