The Convergence of Construction and Computer Science
One topic frequently discussed in JBKnowledge is the Convergence of Construction and Computer Science. As the construction industry moves forward, market leaders continue to integrate elements of computer science into construction science to solve old problems with new solutions. The convergence of construction science and computer science has three primary impacts:
- It gives data scientists access to a data-rich industry ready for data-driven decision-making.
- It gives artificial intelligence some greenfield to solve problems.
- It enables rapid prototyping to prove solutions incredibly quickly.
Deep Diving into Data Lakes
Across all industries, we see a common trend to create data lakes. Firms of all disciplines are aware of the value of “big data,” but lacking in data scientists to interpret it. Rather than immediately investing in data scientists, they are investing in the creating of big data for later use. These firms are rushing to store as much data about their respective industries as possible, so when the data scientists become available, they have a store of data to begin extracting knowledge from. In data science, the massive dump of data is referred to as a data lake.
ENR’s Debra K. Rubin recently published “Data Mining Gains More Cachet in Construction Sector,” an article outlining the state of data science in the construction industry. Rubin’s research clearly shows that market leaders in construction have joined in on the data rush, and are building their own data lakes. Firms are pouring preconstruction and construction management data into their data lakes with the aim of gaining insight into their own operations to increase efficiency.
For example, dumping management data like the movements of workers and equipment on the job site into a data lake gives data scientists data points to evaluate for things like worker productivity, optimal placement of equipment, and equipment productivity and usage that can be extrapolated to predictive analytics of equipment failure or preventative maintenance. As companies continue to build their data lakes, data scientists will be able to leverage those data lakes to capture more and more and then private data-driven recommendations to improve processes, performance, and grow margins.
Building on Artificial Intelligence
Along with data science, artificial intelligence (AI) is an often ill-defined but enormously powerful convergence point of construction and computer science. Artificial intelligence is defined in two-classes: the theoretical strong-AI, and common weak-AI. Strong-AI is the theatrical, self-aware, conscious AI that isn’t yet a part of reality. Weak-AI, conversely, is what is changing all industries today. Weak-AI is simply intelligent, adaptable software designed to problem-solve.
Consider BuildingSP’s GenMEP software. Traditionally, when building MEP components in BIM, modelers use intuition and experience to build systems to the best of their ability, and may use some software tools to analyze what they created and iterate improvements. GenMEP, on the other hand, is an example of construction AI. After the user defines the starts and ends of the pipe runs, GenMEP will automatically generate the pipe runs through the BIM model or point cloud and present the best possible solution. Rather than relying on human intuition, the software itself calculates the most efficient possible way to run pipes, factoring in cost and installation.
This AI uses analysis from the very beginning to generate best-solutions, rather than using retrospective analysis to iteratively help improve an imperfect human-design. AI is already being integrated into BIM software to help builders improve their efficiencies in a traditionally inefficient industry, and is expanding rapidly to automation, robotics, and more.
Rapid Prototyping for the Win
Last, the convergence of construction and computer science is changing the way new solutions are made. Traditional product design is an expensive and laborious process. Whether solutions are software or physical, designs must be defined, built, and tested. The convergence is making this process faster, cheaper, and easier by orders of magnitude.
Several weeks ago, I attended Aggies Invent, an interdisciplinary student event at Texas A&M University that locked students in for a weekend to try to prove solutions to major challenges in construction by building and testing prototypes in a weekend. Students of all disciplines used computer science in nearly every solution. The winning solution was a $1 device aimed to prevent ladder-misplacement accidents. The device adhered to a ladder and provided a loud alarm and vibration to alert a worker that the ladder was not placed to OSHA standards. The students (very few of which were studying computer science) produced this prototype using an off-the-shelf microprocessor and some sensors in less than 72 hours.
These rapid prototyping tools – traditionally reserved for the expert scientist, are becoming available and usable to untrained students with incomplete degrees to understand and implement in a weekend. Companies no longer need to invest enormous resources into research and development, they just need to invest a handful of talented individuals and some time.
The convergence of construction science and computer science has already happened, and the construction industry is at an inflection point. As data-driven decision making becomes more accessible, market leaders will emerge that make their business decisions based on interpretation of huge volumes of raw data rather than intuition. Artificial intelligence will continue to create near-perfect solutions to problems significantly faster than humans. And last, the tools are available to the industry to move quickly, economically, and en-masse.
About the Author
Graham Leslie is the JBKnowledge Research & Development Team Lead (JBKLabs), which is dedicated to disrupting and accelerating the architecture, engineering, and construction industries by building solutions with emerging technology. Graham is a computer scientist with particular research interests in mixed reality and reality scanning. JBKLabs is available for advisory, research, and custom software development services. Learn more at jbknowledge.com/labs.