Want to predict real estate market activity? Chris Morris has an approach for that
April 12, 2019 - Leila Ugincius
It’s unusual for a college junior to easily review and understand research articles in industry journals, said Manoj Thomas, Ph.D., associate professor of information systems at Virginia Commonwealth University. But Chris Morris isn’t your typical undergraduate student.
Before Morris applied for the Undergraduate Research Opportunities Program summer fellowship last year, he took the initiative to review relevant literature and approached Thomas with questions that stemmed from reading the professor’s journal articles on related topics.
“Even though Chris was a [financial technology] major in the VCU School of Business, he had an interest in applying new and emergent data science technologies to advance decision support capabilities in the real estate domain,” said Thomas, who became Morris’ UROP mentor. “It is not the case that a junior undergraduate student would easily understand these research articles. I found him to have the scientific aptitude required to synthesize research questions, the technological competence to develop solutions and the academic focus to effectively complete the research project.”
Morris’ project, “A Pluralistic Approach for Knowledge Discovery in Real Estate,” addresses the lack of decision support in real estate to predict property title exchange. His proposed pluralistic methodology uses various machine learning approaches to construct a predictive model that characterizes real estate market activity. The project utilizes many big data technologies and applies them to data analytics and multicriteria decision analysis, two not-so-easy topics to digest, Thomas said. Morris will then use real estate and municipality data to validate the methodology by predicting when property title exchange is likely to occur.
“It’s just applying math to industry,” Morris explained. “You can apply it to really anything. You can apply it to, of course, financial services and investments. You can apply it to insurance, you can apply it to credit risk modeling. Basically, what we’re trying to do is use data mining techniques or data analytic techniques to find probabilities that certain market activity will occur. So within the research, there are sales prices that you can predict. And there’s also homeowner exchange and the likelihood of a home transferring ownership.
“Within industry or in the research community, there are not [many] data mining techniques that have been used for real estate besides pricing. So if you can apply it to exchanges of home ownership, that’s another aspect of research that the world can use and benefit off of.”
Morris said he chose to concentrate on real estate for this project because it’s a key indicator of the general health of an area’s economy. It is also a sector that can benefit from a pluralistic approach, analyzing data in a variety of ways.
Plus, Morris already had real estate experience. As a freshman, he started his own real estate investment business — Castle Core — with his brother Kenneth. When the brothers parted ways professionally in 2018, Morris wanted to focus on technology-based applications such as those used by financial engineers on Wall Street. “That’s what they do,” he said. “Apply mathematics and computer technology to the stock market.
“Mostly I chose [to apply his model to real estate] because I wanted to align myself in a position that I thought would be the most beneficial,” he said. “If you’re into modeling data, it’s very important to have a domain knowledge of whatever you’re doing. So me having experience in domain knowledge in real estate with my prior experience allowed me to have a better advantage of using the data. … And before I wanted to go into another industry, I wanted to make sure that I didn’t leave real estate behind without trying to apply technology to it.”