The Application of Data Science in Construction Project Management

Machine learning, artificial intelligence, and deep learning are some of those terms that you hear bandied about quite often in this present world, as humans’ reliance on technology grows at an astonishing rate. Though familiar, they remain difficult to comprehend by many, especially for those who can’t make head nor tail of the science surrounding computing technology.

Scouring the internet for articles on technology or having yourself converse with IT experts, you will notice that these big terms are often brought up in the same breath. Although they are bundled together under the same umbrella of advanced computing, these trending buzzwords do not actually refer to the same thing.

Regardless of the differing interpretation, machine learning, artificial intelligence, and deep learning have ushered in a slew of technologies that previously only occurred in science-fiction movies, but they wouldn’t have probably been a reality if there hadn’t existed a new breed of data analysis experts.

Data science is a highly interdisciplinary field of study that involves the extraction and application of meaningful insights from raw data. Many people only associate data science with artificial intelligence, machine learning, and other things correlated to programming. As a matter of fact, it has also been proven extremely useful in enhancing management and decision-making processes in organizations in varying industries, particularly the construction industry.

Machine Learning, Artificial Intelligence, and Deep Learning – What Are They?

For many people who grew up in the 1980s like myself, we were first introduced to artificial intelligence in one of the most influential movies – The Terminator. Up until we actually saw a real, working AI machine with our own eyes, we had been anticipating the day when the earth would be engulfed in flames with multiple swarms of robots moving from place to place eliminating mankind.  Instead, AI machines today have improved our quality of life significantly, holding the golden key to the emergence of the fourth industrial revolution.

Artificial intelligence is merely the wider concept of getting machines to imitate human behaviors. An AI-enhanced machine has the capability to learn and make decisions without any human intervention. Several types of AI machines exist today – we will not cover these in detail, and classifying a machine is based on its likeness to the human mind – i.e. its ability to think and feel like humans.

One of the holy grails of machine learning is the integration of artificial intelligence into machines. Machine learning is an application of AI – you can think of it as a method that translates the concept of AI into reality. It is centered around the idea that we should be giving machines the capability to access data, learn from the data, and make decisions themselves. Deep learning is more advanced than machine learning although their functions are the same. When a machine learning model produces an inaccurate result, someone has to step in and make an adjustment to the algorithm. However, a deep learning algorithm can ascertain on its own if the prediction that it produces is accurate through its neural network.

At the end of the day, these different learning models serve as a mechanical function in the same fashion as a computer, a car, or even a thermometer does. Having said that, here comes the most important subset of artificial intelligence, deep learning, and machine learning, data science.

What is Data Science, and Its Application in The Management of Construction Projects?

A deep learning or basic machine learning model can only be as good as the data that is fed into it – not to forget also the effectiveness of the algorithm. Without data, there is nothing for a machine to analyze and learn. In a nutshell, data science is a field of practice – this means that it can be applied without the use of any technology – while machine learning and deep learning are more like tools and methodologies.

Outside the realm of artificial intelligence, many companies utilize data science as a practical means to extract important insights from data, make a prediction, and suggest a justifiable action. The use of a conventional data science technique – no artificial intelligence or machine learning involved – is only effective in a scenario where the output is only dependent upon one or more parameters that don’t change over time (i.e. f(x,y) = z). In a more complex case where the output depends on a few parameters that shift over time (i.e. f(x(t),y(t)) = z), machine learning becomes handy.f

Many businesses have started to realize the benefits of integrating data science into their management and decision-making processes. Particularly, the construction industry is seeing a large increase in the number of companies that have adopted data science to better manage project deliverables.

Unarguably, poor planning, cost and budget overruns, and disputes are just some of the common problems that are deeply rooted in the industry. For many construction companies, avoiding these problems seems nearly impossible. However, with the increasing application of data science, they may one day become a thing of the past.

Speaking of that, we will explore some of the ways that data science is improving construction processes.

Predictive Analytics

Predictive analytics is actually an important, well-known analytical tool that businesses – especially ones involved in construction – use to make justifiable predictions on future unknown events. It analyses past and current data and, through machine learning and/or manual calculations, makes a prediction about the future related to the data.

It is very common for construction projects to face various uncertainties – these are commonly known as risks in the construction context, on a daily basis. Because of that, this technique is certainly a real game-changer that can help predict the rise of a particular risk before it even occurs.

Financial Management

Using data science, construction companies can create an accurate yearly budget plan that takes into consideration the usual peak and off-peak periods for their business activities. Based on past data, managers are able to accurately determine when the companies will be the busiest and develop a budget plan for the business accordingly. This is relatively useful in a seasonal industry like construction, as an off-peak period can really put a company out of business.

Health and Safety  Management

Basic site sensors can be placed around a construction site to gather environmental data related to on-site temperature, noise intensity, and dust particulates. Data collected will assist in monitoring workers’ exposure to these environmental hazards and allow the early identification of issues before they arise – issues that may impact the health and safety of workers.

Data science also helps prevent worksite-related accidents; construction sites are considered to be among the most hazardous and accident-prone work environments. Using historical data, it is possible to predict the probability of an accident occurring at a particular site. By comparing a job site-working condition and the work ethics of the workers with the data that correlates with these two areas, a risk rating – the risk of an accident occurring – can be assigned to the project. A safety briefing can promptly be held upon the detection of a threat before the matter goes out of hand.

Material Management

Finally, yet importantly, effective material management plays a huge role in the outcome of a construction project. With data science, managers more accurately forecast how much of certain inputs they will need for their assigned jobs. For large and complex projects, something as simple as overestimating the number of bolts or nuts required can potentially cause the budget to go overboard.

Contact Leopard Project Controls for your next CPM Scheduling Project.

Let us look at a very simple scenario in which a job involves the placement of formwork that requires a certain number of fasteners per square foot. And, we have gathered some data that shows for that particular type of formwork system, the number of wasted fasteners per square foot follows a concave-down parabolic pattern for some reason. Using data science, we can manually find the line of best fit for the data (i.e. f(x) = y, where x is the number of square feet) and come up with the exact quantity of fasteners required or a better and more economical formwork arrangement.