Due to a variety of factors including the size and scope of construction projects, the volatile nature of the workforce, and the conflicting interests of the different parties involved, construction has always been a complex industry where investors, developers, and those who design and construct are prone to a variety of operational, financial, and legal risks. A comprehensive monitoring program that integrates data analytics into site monitoring and compliance can assist stakeholders in developing and implementing proactive and reactive protocols in order to identify, measure, and mitigate risks around construction project management. This article discusses a case study where a multi-billion-dollar construction project benefited from data analytics to monitor more than one thousand construction workers from nine different contractors, in order to detect and prevent fraud, waste, and abuse.
A prominent estate firm was working with numerous contractors when construction started for the largest private development in the United States. The legal counsel of the real estate development firm was tasked with monitoring the contractor companies and their unionized construction workers in order to detect and deter fraud, waste, and abuse on the construction site and prevent violations of laws, regulations, policies, and procedures. The legal counsel enlisted the help of specialist consultants to assist with the data and technology aspects of the site monitoring.
The engagement required the collection, preservation, and processing of data from disparate data sets, which included:
- attendance data from daily attendance sheets prepared by team leads on the construction site,
- labor distribution reports, which are accounting statements showing hours worked and wages earned,
- periodic electronic payroll exports from subcontractors, site access data from the security software managing access rights to the construction site, and,
- video footage from the surveillance system.
One of the challenges faced by the construction industry is the record keeping discrepancies between developers, contractors, and project management. Obtaining and analyzing data from disparate data sources ranging from attendance sheets, to payroll data, to the time stamps of the card swipes of construction workers entering and leaving the construction site can be useful in bringing all of the data sources together and identifying any discrepancies in reporting that may lead to losses for the project as a whole. This reconciliation exercise also helps to identify any suspicious or anomalous activity and inform key stakeholders of any undesirable behavioral trends that may be arising on the site.
Attendance sheets show which workers were present on the construction site for a given shift. Labor distribution reports show the total straight time, overtime, and double time hours paid to each worker. A high-level reconciliation can help identify any discrepancies between timesheets and payroll, and pinpoint any instances of overbilling by individual workers, contractor companies, or labor unions.
For this multi-billion-dollar project, the forensic data analytics team was able to take payroll reconciliation one step further by integrating site access data from the physical security software that operated the turnstiles at the numerous gates of the construction site.
In order to enter and leave the construction site, construction workers swipe their ID badges at the turnstiles. The security software recognizes their credentials and grants them access. The main purpose of this infrastructure is to prevent unauthorized parties from obtaining physical access to the construction site. However, the security software does more than that: it also collects time stamps for each individual entering and leaving the site, and this data is collected and kept in a database.
For this engagement, the data analytics team utilized the time stamp data from the swipe activity to calculate the number of hours worked by each construction worker, considering lunch and other breaks taken throughout a typical shift. The automated calculations of hours worked provided an additional data point to compare against the hours reported in timesheets and labor distribution reports. The discrepancies between the daily swipe calculations and the hours reported on the labor distribution reports were used to identify individuals who had overbilled time on any given day. For instance, reconciliation of swipe data against timesheets identified workers who were marked as present on days when they did not have any swipe activity.
Of course, as anyone who has dipped their foot in a data lake will tell you, there are always anomalies in the data. It could be the case that a worker happened the forget their ID badge, or that the attendance taker was not paying attention on that day. These are plausible scenarios, and it is the job of the data analytics team to look into anomalies and distinguish between anomalies that are caused by missing or incomplete data and those that potentially signal fraud, waste, or abuse.
When timestamps from swipe activity are being used to calculate working hours, the expected pattern is that a worker arrives at the construction site at the beginning of their shift, takes a lunch break midway through their shift, and occasionally takes additional, shorter breaks throughout the day. Based on this understanding, swipe data was used to detect individuals who take longer breaks, as well as identify contractors whose employees have higher averages of break times. Furthermore, it was discovered that some workers did not return to the job site after their lunch break.
Further analysis of the timestamps also led to the discovery of another type of anomaly, where some workers were ”double swiping” as they entered the worksite, where they were swiping an additional ID badge belonging to another worker after going through the turnstile themselves. This pattern was first discovered by detecting pairs of ID badges whose time stamps appeared too close to each other. The surveillance video footage was then used to confirm the double swiping activity.
The data analytics team produced an interactive Site Overview dashboard, which automatically updated as the data feeds refreshed. The dashboard consisted of contractor metrics to help the real estate developer determine how the different contractors on the construction site were performing in terms of productivity and efficiency.
The most obvious and perhaps most directly relevant outcome of the site overview reports is the ability to examine the ways in which the contractors differ from each other in terms of attendance, timesheet and payroll reconciliation. The real estate developer was able to take action in addressing issues with contractors whose calculated worksite hours fell short of the number of hours reported on the labor distribution reports.
An additional benefit of the Site Overview dashboard was to be able to track trends over time for each individual contractor. For instance, the real estate developer found that their biggest contractor, while increasing in headcount, was showing a decrease in the average number of hours worked by each worker, which signaled a decline in productivity and efficiency. The same contractor also showed workers becoming less compliant with swiping into the worksite by using their ID badges, which was corroborated with surveillance videos showing the workers jumping the turnstiles. The data also showed additional patterns, such as the work hours on Fridays decreasing over time.
In order to enforce compliance by individual workers, daily reports were issued to highlight problematic behavior that can be immediately addressed on a case-by-case basis. The anomaly detection metrics produced a mid-day report that listed the workers who had breaks that lasted for longer than one hour, and those who did not swipe back into the job site after their lunch break. The end-of-week report produced a higher-level view into the anomaly metrics, highlighting contractors whose workers were taking longer breaks on average.
Value-Added of Data Analytics
Implementing a data-driven approach can help companies increase control over their construction sites by collecting, integrating and analyzing data on their projects and contractors, and ultimately create an integrated platform where they can reconcile third-party data with in-house data, integrate sources, and highlight any discrepancies, inefficiencies and noncompliance. By integrating the actionable insights obtained from data analytics into the decision-making process, stakeholders can achieve better access to information, faster response times, reduced costs, and advanced reporting capabilities.
Future of Site Monitoring
As further technological advances are made, we uncover not only additional data points but also additional ways of obtaining existing data points. For instance, in the project described above, manually prepared timesheets collected from team leaders had to be transcribed into electronic format prior to reconciliation. This process can be streamlined by the introduction of an electronic attendance taking system.
The whereabouts of the construction workers were calculated using the time stamps from the swipe activity at the turnstiles. This allows for determining when the worker is within the boundaries of the construction site but does not precisely measure the level of activity (or idleness) within the site. By integrating mobile applications installed on smart watches or smart phones with barometric pressure sensors, the exact coordinates of the workers could be known. More specifically, the activity and productivity of higher paid construction workers such as crane operators can be monitored by examining their altitude within the site.
One step further than commercially available smart wearables is a construction-specific smart wearable, which is the smart hard hat. The construction industry has already been experimenting with attaching motion sensors to hard hats to keep track of the workers on construction sites. In the recent years, start-up companies have developed smart hard hats (1) that are capable of doing much more: They track the proximity of the workers to each other to make sure they are properly spaced for safety and efficiency purposes, they alert workers that may be falling asleep by detecting the ‘nodding off’ motion of their heads, and the more advanced models can even superimpose instructions and augmented reality onto construction sites for an enhanced view of the project. These devices have become even more desirable with the COVID-19 pandemic, since the technology can help with both social distancing and contact tracing.(2)
As more data points become more readily available, there will be opportunities for wider applications of data analytics to construction site monitoring. With any luck, we will be able use data analytics to minimize not just financial losses but also safety incidents, worksite accidents, and even the spread of a global pandemic.
 The Field lens Team. “Move over, smartwatch: The smart hard hat is here”. Field lens Industry News, 26 July 2019, https://fieldlens.com/blog/building-better/smart-hard-hat/, accessed 5 November 2020.
 Mortice, Zach. “A Smart Hard Hat Designed to Keep Workers Awake Can Also Trace COVID-19”. Redshift by Autodesk, 16 June 2020, https://redshift.autodesk.com/smart-hard-hat/, accessed 5 November 2020.
© Copyright 2020. The views expressed herein are those of the author(s) and not necessarily the views of Ankura Consulting Group, LLC., its management, its subsidiaries, its affiliates, or its other professionals. Ankura is not a law firm and cannot provide legal advice.