Dive Brief:

• Suffolk Construction Co., after hiring a chief data officer and a team of data scientists last year, said it is developing artificial intelligence tools that can analyze photos from jobsites to identify risks and prevent worker injuries, according to MIT Technology Review. 
• The Boston, Massachusetts-based general contractor said its data group it is also developing a computer script that can look at information the firm has collected, such as a decade’s worth of scheduling data, to forecast project delays.
• The new digital tools could help Suffolk increase productivity by 14 to 20% in a few years, according to Jit Kee Chin, the company’s CDO. The C-suite role, dedicated to “leveraging big data and advanced analytics to improve the core business,” according to the company, is unique in the construction industry, which often carves out lower-tiered positions to managing data projects. A 2017 McKinsey report went further, saying that construction firms could boost productivity by as much as 50% through data analysis.

Dive Insight:

Suffolk, No. 26 on Engineering-News Record’s Top 400 list of contractors by 2017 revenue, has invested in other big data-driven projects, such as its Smart Labs facilities, a place in which it “will identify, test and scale new technologies intended to transform the construction experience and revolutionize the industry.”

One facility, for example, features its Computer Aided Virtual Environment (CAVE), an element that allows clients to immerse themselves in a 3D representation of a project via nine screens along four walls of connected smart boards that collectively measure project progress in a visual flow chart.

The firm may be farther along than some of its competitors, but the industry is catching up, slowly but surely, according to software and consulting firm JBKnowledge’s 2017 Construction Technology Report, which was referred to in the MIT Technology Review article. CEO James Benham said that around 20 construction firms active in the U.S. have embarked on data science initiatives in recent years.

Many, however, haven’t yet applied the tools to safety, like Suffolk’s new push to write algorithms that detect safety risks. The new technology reportedly scans 700,000 images from 360 projects during the past 10 years through the lens of image recognition software to detect workers’ use of hard hats, gloves and safety vests and goggles. The data team, led by Chin, reportedly then plugs the analysis into other sets of data related to weather, project timelines and more, and a second machine-learning model predicts delays and makes forecasts on current and future projects.