Georgia Tech's Urban Analytics Degree Seeks To Train Cities' Future Problem Solvers
Think about your biggest day-to-day roadway headaches like traffic, finding a parking space or avoiding that intersection with the eternal red light. Now, imagine a computer solving those problems.
That’s the goal of a new degree program launching this month at Georgia Tech. The Master of Science in Urban Analytics combines the fields of urban planning, computing, and industrial and systems engineering to “fix big city problems.”
It’s not just about transportation. Urban analytics could be used to prevent crime, mitigate flooding, limit air pollution, keep housing prices affordable — the list goes on.
“Urban analytics is essentially a discipline that uses data and data science tools to solve urban problems,” said Subhrajit "Subhro" Guhathakurta, professor at Georgia Tech’s School of City and Regional Planning.
He’s an urban planner who invited engineers and computer scientists to team up to create the new degree program.
GPB’s Rickey Bevington spoke with Guhathakurta about the program’s debut.
Subhro Guhathakurta: One example that comes to mind right away here in the Atlanta metropolitan area is the use of Peach Pass, where the price of what travelers pay along the Peach Pass is determined according to the demand for that lane at that specific time. And that is gauged by data. As more and more cars enter the freeway, the cost of using the lane goes up. And as the demand is lower — fewer people using the lane — the cost goes down.
Rickey Bevington: Beyond traffic, what other big problems could your students be solving and why focus on urban issues?
Subhro Guhathakurta: Urban areas have enormous challenges that a lot of these cities are facing across the world. And these challenges include issues of segregation, of housing affordability, of energy efficiency, of water demand. There are a host of issues: Pollution, climate change issues. And given that 60% of the world is now living in urban areas, these problems are going to be much more acute for us to resolve.
One of the ways we can resolve these problems is by using data and having good information about how people live their lives so that we can better understand their lives and livelihoods and solve the bottlenecks using data science and better information.
Rickey Bevington: Where does all this data come from?
Subhro Guhathakurta: Well, data comes from a variety of sources. Even the mobile phones that we are using all the time is a wonderful source of data. There are sensors that are used across the city which are sensing things such as levels of pollution, levels of noise, heat, traffic counts, pedestrian counts. So a lot of sensors are used across the city to provide us with important information on a minute-by-minute basis.
The question is, how do we use that information to make better decisions? And if we can use that information to make better decisions, we can perhaps find ways to automate decision-making in some simple ways, but also be very conscious of the fact that automating decisions may lead to certain kinds of biases or errors in judgment. So we have to be also very careful about what parts of our decision-making process we can automate, and what we need to control and have a human in the loop to make those decisions.
Rickey Bevington: We can't have a conversation about data without talking about ethics and privacy. What are you going to be teaching these students about those issues involving data for the next generation?
Subhro Guhathakurta: Ethics and privacy were a very critical part of the degree program as it was designed. These biases have been well known. For example, we know that automating issues such as mortgage lending, for example, have led to biases where certain groups of people would be considered to be more risky, even though considering the data specifically of that group does not suggest it. But data science approaches might not be looking for these kinds of biases. We have to be careful about understanding biases so that we don't fall into the trap of making decisions that are unfair or unjust or hurtful to certain groups of people.
And sanitizing data is something that is very important to make sure that we are not violating the privacy of individuals. So to sanitize that information, we actually figure out ways to create certain types of “noise” in the data so that individuals are not identified directly. But when we aggregate data, it still provides us with useful information that we can act upon and make good decisions with.
Rickey Bevington: This master's degree program is getting underway for the first time in the next couple of weeks. How many students do you have participating right now?
Subhro Guhathakurta: The first class that we have has 22 students in that class. But in terms of the degree program, we have about only four students who were able to apply given that our application process was all of three weeks. But we expect that in the next round of applications, we will have a lot more interest from various students, from different disciplines.
Rickey Bevington: I think that any young person would be energized at the idea of learning the skills to solve the planet's biggest problems. So congratulations on launching this new program. And thank you for joining me.
Subhro Guhathakurta: Thank you for having me. It was a pleasure speaking with you.