Â鶹´«Ã½

The US Occupational Safety and Health Administration says loud noise is an enemy of productivity. Often, it interferes with communication and concentration.

This can lead to complaints and even potential legal ramifications for our clients. Traditional noise monitoring methods include using sound level meters in the field to measure noise levels from construction.

When a preset trigger noise level is reached, it is recorded by the sound level meter. These meters capture a short audio clip that is associated with the triggered noise level. And we can identify the source of the noise.

The current industry approach is for a staff member to review each audio file and mark it as ¡°construction¡± or ¡°not construction.¡± That helps clients identify if a noise level exceedance is from their project.

For large, multiyear projects, employees may need to review and classify thousands of files. The turnaround time can be hours or days. And that means the construction noise may have already impacted the community or exceeded environmental noise conditions.

Our team has developed a construction noise audio-classification tool that uses the power of artificial intelligence to categorize these audio clips. This eliminates the staffing requirement and speeds the response time.

How did we pull this off? With data.

We trained the tool on with thousands of audio files, so it could learn the difference between construction noise and nonconstruction noise. During the testing phase, the tool achieved an accuracy reading of over 99 percent and could process 1,000 audio files in less than 15 minutes.

We are coordinating the use of the tool for ongoing construction projects. Data science and digital technology made this possible.

2.?Digital science and data can bolster flood science

In 2022, catastrophic flooding hit the eastern part of Kentucky. Forty-five people were killed, and thousands were displaced. Severe rain events like this are becoming more common with our changing climate.

Communities need better ways to mitigate floods. They need to understand what areas are more prone to flooding than others. Using traditional modeling methods, that¡¯s an effort that could take years.

However, our Flood Predictor digital solution slashes that timeframe. We¡¯ve used data¡ª30,000 hours of engineering flood models¡ªthat is augmented with theory and Â鶹´«Ã½ to analyze a large part of eastern Kentucky in just a few days.

We¡¯ve signed an agreement with the Kentucky Division of Water to use Flood Predictor to provide flood risk mapping, assessment, and planning services across the state. The volume of data¡ªplus robust training, validation, and testing of our machine learning algorithms¡ªhas made delivering such data-driven insights possible.

The state of Tennessee has also experienced extreme weather events at an increasing frequency. It has caused severe flooding, destruction, and the tragic loss of lives.?

The Tennessee Economic and Community Development Department has integrated Flood Predictor into the TNPlan web portal. It¡¯s a data tool that gives community officials and emergency managers quick access to flood risk insights¡ªbefore severe weather hits.

We¡¯ve also worked with Walton Street Capital, a private equity real estate investment firm, to provide data-informed flood risk decisions for real estate locations across the US. We used Flood Predictor to prepare a flood risk analysis for a number of properties within the firm¡¯s portfolio.

This commitment to digital science and data, plus our subject matter Â鶹´«Ã½, will work to protect communities against severe weather in the future.

3.?Data-driven dam safety

We have a mining client in Western Canada that utilizes tailing dams in their operations. Tailing dams store water and waste by-products from the mining process. The material could cause serious environmental damage and financial loss if the dam was breached because of erosion, flooding, or geohazards.

Across the globe, there are about 3,500 tailing dams. According to a 2019 study, 10 percent of those have experienced stability issues at some point.

It¡¯s a big problem.

Attentive and efficient monitoring is essential to maintaining the health of tailing dams. How do we help our clients achieve that goal? With our Stantec Mosaic??digital solution.

The Dam Insights tool aggregates vital data from instruments situated along the dam to provide a snapshot of the dam¡¯s health. Typically, staff members would have to analyze the instrumentation data and determine the risk factors facing the dam.

That process can be time consuming. And, depending on the severity of the issue and location of the instruments, the recognition of a potential failure made by a staff member could be too late.

The differentiator with Dam Insights is that the tool is trained to analyze the data, removing the human element. It will provide reports on each piece of instrumentation and let dam owners and operators know which area of the dam requires attention.

That saves time. And it helps the client make decisions according to perceived risks facing their assets.

Data will lead the way

It¡¯s obvious that data and digital technology can help solve problems faster and more accurately across the AEC industry. That¡¯s important because financial constraints are real, and clients want to reduce costs but still provide needed services.

This means making a further push into the world of data science and digital technology is a prerequisite for future success in the AEC industry.

Data-informed decision-making is the way of the future. And the future is now.