Untitled document edited 252

Cabello, Martínez-Rojas, Carrillo-Castrillo, and Rubio-Romero authored the paper titled
“Occupational accident analysis according to professionals of different construction phases using
association rules” and had it published in the Safety Science journal in 2021. The paper
addresses the problem of construction safety because of the perpetually large numbers of
avoidable accidents at construction sides. By employing the association rule data mining method
to extract data about accidents from achieved records held by repositories in different countries
in Spain, the authors created a framework for improving construction safety practices.
Identifying the phases of the construction sites at which specific accidents occur enables the
development of better accident prevention strategies.
Summary
This paper aims at developing a strategy for improving accident prevention strategies in
construction sites, which are deemed to have the highest frequency of occupational accidents
across different sectors. Although the paper uses secondary data already collected by various
construction companies in Spain between 2003 and 2015, it uses the data to explore the factors
that contribute to occupational accidents at construction sites in a new light to reveal novel
relationships. After mining and filtering the data from construction company digital data
repositories in Spain, the authors end up with 1,525,865 incidents related to construction safety
for consideration (Cabello, et al. 6). After that, using different categories and parameters
pertaining to construction accident occurrence, the authors employ association rules, a renowned
machine learning technique, to unearth hidden relationships between the variables in the
construction safety dataset. The safety risks at the construction sites are associated with the stage
of construction, which the authors categorized as i) the foundation and structure phase, ii) the
facilities and masonry phase, iii) the completion phase, and iv) the phase described as
encompassing the construction process (Cabello, et al. 6). Their correlation analysis involves 58
variables.
Critical Analysis
The introduction of construction phases as variables in analyzing construction site accidents is a
unique feature in the study conducted by Cabello et al. This study does more than identify the
causes of construction site accidents to how these causes vary at different construction phases.
The authors also introduce employee demographics, company details, and other accident-related
details, such as the day of the week when the construction accident occurred. This approach of
variable definition granulates the causative relationships between construction accidents and
worker, workplace, and environmental factors at the construction site, which is unique in this
kind of literature.
The focus of the study leverages new trends in research conduct, including the use of technology.
Specifically, the authors use digital data from construction companies, indicating that they are
conversant with the technological advancements in the construction industry, and particularly the
use of information management systems in construction sites. The digital repositories hold
enormous data that harbor intricate and subtle relationships between diverse variables associated
with construction worker safety that are often difficult to decipher. To overcome this hurdle, the
authors employ association rules to unearth the hidden relationships between the variables
related to the construction personnel, accident, firm, and project, thus helping to discover the
association between the distinct components of the same occurrence. In this case, the authors use
algorithms, such as the Apriori algorithm, to discover Boolean association, alongside several
others.
Although the technical paper deals with a complex subject, the authors present their information
in a coherent manner that can be followed by a reader with limited technical construction
knowledge. Apart from the generous use of diagrammatic representation of complex concepts
and data, the authors endeavor to explain these concepts before applying them in their data
description and analysis. Similarly, the methodology employed in the study is sufficiently
detailed to enable repetition of the study by a different researcher. However, the authors avoid
several technical details that leave a knowledgeable reader dissatisfied. For instance, the author
does not disclose all the association rules employed on the data set, which appear to exceed 450
(Cabello, et al. 8). However, this is understandable because disclosing every association rule
used in the analysis would require enormous authoring space, which is unavailable in such a
technical paper. Besides, the authors use many references throughout their entire paper, which
can be used by readers that are curious about additional information on the subject. This not only
places this article amongst other peer-reviewed ones to enhance its credibility and validity, it also
demonstrates the ethical grounding of the authors when presenting technical information related
to a sensitive subject, such as construction safety.
Conclusion
This article deals with problematic issues of construction safety with elusive solutions. The
authors recognize that safety in construction sites is hampered by the poor understanding of the
causative factors and their association with the various construction site conditions. The authors
take advantage of technology by using machine learning techniques to mine data and associate
different variables related to the worker, the construction company, the construction accident and
the construction project. The authors have improved the decision-making process for developing
accident-prevention strategies in construction sites that are more targeted to reducing specific
and frequent hazards and risks.
…