Like any city, Boston is ever-changing. Old buildings come down, new ones go up. Public transit snakes through the city. Everything from schools to grocery stores to trees dot the landscape.

For years, the city has wanted to capture all of this in a digital 3D model that could help guide planning decisions before they take effect in the real world. That model came to be in 2015, and three years later, the city made it accessible to the public.

This digital twin encompasses the current landscape of buildings, transit, trees, daylight and shadows, and points of interest. It also contains proposed and under construction buildings and serves as an archive of what Boston looks like through the years.

"This is like a living, breathing kind of content in that it has the archives in it, it has current and then it has future," said Carolyn Bennett, deputy director of GIS, or geographic information system mapping, for the Boston Planning and Development Agency. "It sort of covers the whole gamut, and I think that that's helpful for the agency and the city of Boston as a whole."

Digital twins, 3D virtual replicas of a given system, place or thing, allow cities and property owners to test out changes before they enact them in the real world. Cities from Los Angeles to coastal Texas and beyond are increasingly using this technology to study the effects of development, traffic, climate change and myriad other situations a city must face.

"You can simulate the effects of your decisions on a physical object, in this case a city or a campus, before you actually make those changes in the real world," said Ankit Srivastava, associate professor of mechanical and aerospace engineering at Illinois Tech. "It makes the decisions much more data-driven and much less expensive because you can play with various possibilities."

But it takes dedicated time and effort to ensure that a digital twin remains up to date and helps a city move forward. 

Capturing a city in the digital realm

Since 2005, Boston has been having conversations about how to incorporate the city’s troves of data in a 3D model, Bennett said. Available data included geospatial data for water and sewer, transportation, tax parcels and more.

Now, Boston’s digital twin contains that library of data. The city can use it to look at development proposals and their effects on a neighborhood in terms of housing, zoning and parking. Planners can take urban heat island data from other sources and overlay it onto the twin to visualize temperatures in reference to buildings, impervious surfaces and tree canopies, for example.

"It's a better way to sort of visualize what we do as planners," Bennett said. "We plan in 3D. We're not a 2D agency. We look at the world in its existing form, so you kind of have to look at it in 3D."

Chattanooga, Tennessee, has also found a digital twin to be a useful tool to encompass the data it collects.

"Can we bring all this real-time data into one common ecosystem, where we can get a really good picture of what is happening?" said Jibo Sanyal, who leads the computational urban sciences research unit at Oak Ridge National Laboratory.

Researchers at Oak Ridge and the National Renewable Energy Laboratory partnered with Chattanooga to build a digital twin that helps anticipate and alleviate traffic congestion. Sanyal is the technical and strategic lead on the project.

Information from 500 different sources, such as traffic cameras, 911 data, radar detectors and weather stations, feeds into Chattanooga’s digital twin. Traffic congestion experiments done within this virtual realm have shown an improvement of up to 30% in traffic flow, resulting in greater energy efficiency, Sanyal said.

Researchers use the findings from these digital twin experiments to enact changes in the real world. For example, transportation planners typically collect data during the morning and afternoon traffic peaks. But when researchers ran digital twin experiments on Shallowford Road in Chattanooga, they found that more than 90% of cars get stopped by red lights during the midday rush.

In the real world, these researchers have used the insight from those experiments to recompute the timing for traffic signals along Shallowford Road to reduce congestion. They now can change their signaling pattern every four minutes based on traffic conditions, Sanyal said.

Chattanooga has one of the country’s busiest freight corridors, so the digital twin also allows for longer-term decisions. For example, the city wanted to know, are there certain times of day when it makes sense for drivers to use the shoulder as a travel lane? Using a digital twin for this work reduces the assessment costs, Sanyal said.

"A lot of research really stops [when] they publish a nice paper, often a very compelling paper, but the opportunity to actually implement it in the real world often isn't there,” he said. “Chattanooga is amazing. They actually let us go in and work with them to change settings in their traffic lights, configure them slightly differently if needed, and run these experiments in the real world. And then we observe before and after to make a determination of what kind of improvements we're seeing."

Though researchers only control signals along one corridor currently, Sanyal said he expects they will control about 100 signals within a year, across downtown Chattanooga and other high-traffic areas of the city.

The buildings that make up a city

Digital twins can also help cities and property owners reduce energy consumption.

Cities produce around 75% of the world's greenhouse gas emissions. In order to reduce this pollution, cities such as New York, Boston and Washington, D.C., have created targets for a future where buildings have net-zero carbon emissions.

But decarbonizing a building requires analysis and implementation of energy management systems and renewable strategies, in addition to the purchase of carbon offsets. Cityzenith is one of the companies that focus on developing urban digital twins that help do just that.

The model can start with as few as two to five buildings, said CEO Michael Jansen. Property owners input their data using a template, and the resulting digital twin helps them run and optimize simulations based on factors such as a building’s age, condition and purpose. The goal is to have more energy-efficient buildings and therefore cities.

Las Vegas and New York City have adopted this technology, with Phoenix and other U.S. and European cities expected to follow. Cityzenith plans to donate the technology to a total of 10 cities by Christmas this year and a total of 100 cities by 2024, Jansen said.

Roadblocks to digital twins

One of the main roadblocks to implementing digital twins is making sure all the necessary data is in place to build out the model, Jansen said. "Digital twins aren’t bought," he said. "They’re built."

Because of the amount of data large cities already have, it is expected that more of them will use digital twins, said Mohammad Heidarinejad, assistant professor of architectural engineering at Illinois Tech. He added that in five to 10 years, these models could make their way to more smaller cities, too.

But cities and property owners will face hurdles along the way. Capturing reality with a drone or scanning requires a lot of storage, plus a digital twin must evolve with time as things change in the real world, Heidarinejad said. (Boston updates its model twice a year.)

One way to update a digital twin is to crowdsource data, including photos that people take of a city and agree to contribute to the project, said Srivastava, of Illinois Tech. Perhaps a central system could analyze those photos and their GPS coordinates and automatically update the digital twin with that information, he suggested.

The data digital twins contain can come with privacy issues. If people can access the information behind a twin, so can hackers. Srivastava said a city must decide how much detail to capture in its digital twin and what elements to make public, especially as the amount of data in the world increases.

"There's going to be more and more virtualization of the real world, and it's only because it's just so much easier to get data," he said. "I think it’s going to be leveraged in the future by cities."

That will likely require some investment from cities. Bennett said Boston has a 3D data analyst on staff whose main responsibility is the city’s digital twin.

"Cities have to be sort of committed to it because it is a living content and data that you have to keep updated and maintained,” Bennett said. “It's challenging in that way because if it becomes too out of date, then people really won't want to use it."

Sanyal of Oak Ridge National Laboratory said the staff and IT overhead can be difficult for a city to take on. After all, cities want to be good stewards of taxpayer money, so they tend to use things that are proven.

The digital twin work Oak Ridge is doing with Chattanooga is designed to help that. Oak Ridge focuses on solutions that are meant to scale to other cities, so Sanyal wants to see companies license the national lab’s technology.

Still, a digital twin is just one tool that cities and properties can use to chart their future, and the value it can provide will depend largely on the data it has to work with.

"Because it is so based on data, it is also very important to be careful of what data we're bringing in to make these decisions," Sanyal said. "Ultimately, a digital twin is a tool, and large-scale planning decisions are usually made by humans."