Monday, November 5, 2018
As an international student from Bangladesh, Mohammad Rahman graduated with a degree in computer science shortly after the Sept. 11 terrorist attacks.
“Because of my name and the fact I was an F-1 international student, it was difficult to find a job,” says Rahman, now an associate professor in the Krannert School of Management at Purdue University and one of the 40 best professors under 40, according to a 2017 list by Poets and Quants. “I was reflecting on what to do, and that’s when I figured, I’ve got to continue my education. That’s the only path that seems to be less resistive than the others.”
The seemingly unfortunate twist of fate would shape the rest of his career. Rahman became an MBA student with a computer science background who, in the emergence of a digital renaissance, foresaw technology’s transformative role in commerce and its destructive potential to obsolete businesses, just as traditional, brick-and-mortar retailers were seeing the emergence of online competitors in Ebay and Amazon, among others.
Now, he studies the myriad ways in which digital platforms have grown to revolutionize the world of retail and fundamentally change how businesses interact with consumers, for better or for worse.
“We’re talking about 2002 and 2003, when new technologies were popping up every day – Amazon, Ebay, Napster – that were changing the way people live, the way they see and do things,” Rahman says. “Today, we are seeing the disappearance of historic stores like Sears and Toys R Us. This is really where I started digging into the whole idea of digital transformations: How omnichannel shopping, big data, and digital traces are changing consumers and their behavior, as well as market structures and business models.”
Rahman’s work aligns with Purdue's Giant Leaps celebration, acknowledging the university’s global advancements made in artificial intelligence as part of Purdue’s 150th anniversary. This is one of the four themes of the yearlong celebration’s Ideas Festival, designed to showcase Purdue as an intellectual center solving real-world issues.
Today, there’s no shortage of material to study among the troves of information produced by companies and customers. But in the dawn of the era of big data, Rahman was among the first researchers to embrace, rather than avoid, overwhelming data sets.
“When I was doing my Ph.D., this whole idea of gathering a huge amount of data and leveraging economic models to generate insights was very rare; now, we call it ‘big data,’” he says. “When I got server logs from a company that were millions and millions of rows, most people would just flee away from it. For me, it was interesting because I came from that computer science background.”
Take, for example, his recent study on Airbnb’s economic impact in neighborhoods in New York and five other cities. Rahman and Ph.D. student Mohammed Alyakoob used web crawlers to collect millions of data points, including 34 million Yelp reviews, and used complex algorithms as well as econometrics to quantify the impact of Airbnb rentals on surrounding restaurant employment.
“Today, when it comes to data complexity, nothing really scares me,” he says. “It’s rather an opportunity to generate simple and useful insights.”
Many social scientists, however, fail to look beyond the data and rely too heavily on preconceived notions, Rahman said. As a result, a focal point of his research has been to pound the pavement and interact with the people he studies.
“One thing that has been important for me is really understanding the context of my research,” he says. “I think this is where you build acumen, you make sure you’re not really missing something obvious. By relying solely on data, you’re basically boxing yourself in with your preconceived ideas. We all have some expectations about our research problems, but that can change if you go talk to the people who are in the field.”
Rahman knows better than others that technology sometimes produces an outcome opposite of the desired effect. One of his biggest fears is that humans’ growing reliance on big data and automation will exploit our biases rather than help us overcome them.
“Digital assistants are going to have a lot of say in influencing the purchases people make, and the data you feed into AI is going to dictate the kind of suggestions these systems are going to offer to people,” he says. “If there is a lot of bias in training these AI solutions, we’re going to see a lot of bias in terms of consumption. A lot of these biases we have – the neighborhoods where we choose to eat or stop for gas – could be systematically injected into our preferences in a much bolder way than today.”