As the fourth largest city in the United States, Houston has a broad real estate market shaped by a complex interplay of financial and economic factors. Understanding and predicting market indicators is of great interest to investors, homeowners and buyers alike.
Over the past five years, real estate trends have been shaped by the COVID-19 pandemic, supply chain fluctuations, inflation, infrastructural changes, and resilience planning efforts to adjust to climate change.
Jian Ruan, a Rice University undergraduate student majoring in statistics, studied evolving trends in Houston’s real estate market from 2017-2024 and presented her findings as part of her senior capstone project.
Ruan’s multifaceted data-driven analysis encompassed geographic information systems (GIS) visualization to explore the spatial dynamics and diversity of new developments, time series analysis to examine temporal trends of new development applications, and machine learning techniques to predict future urban growth trajectories.
Ruan was awarded the Best Capstone Presentation for STAT 450 this spring, and her paper was selected for publication in the CoFES White Paper Series. She also shared her findings during the student poster session at the 11th Eubank Conference on Real World Markets: Real Estate Investment in a Changing Market.
Ruan’s capstone paper presents two data sets for analysis. The primary dataset is new development data obtained from the Houston Planning & Development Department and contains 20,000 records spread over 31 variables. To align with the research objectives, seven variables were selected for further analysis based on their relevance to home values, as recommended by real estate industry professionals.
The secondary dataset, provided by the Harris Central Appraisal District, offers a detailed snapshot of home appraisals and market values in Houston, featuring 900,000 observations of 72 variables. For the purposes of Ruan’s study, nine variables deemed most relevant to new development characteristics were selected. Shared variables like year, zip code, and acreage link the primary and secondary datasets, facilitating integrated analysis.
Ruan's research revealed several key findings about Houston's real estate market. Her GIS visualization demonstrated a shift in the city's expansion pattern from circular to triangular, with COVID-19 serving as a pivotal transition period.
Time series analysis indicated that new development counts are highly sensitive to market shifts, exhibiting a bi-yearly cycle influenced by federal interest rate trends, and forecasting predicts a decrease in new developments in the coming months.
Utilizing machine learning, Ruan identified that market segmentation aligns with major highways through K-means clustering. Both linear and tree-based models highlighted that single-family homes have the most significant positive impact on nearby home appraisal values, with other critical variables including acreage, the number of lots, and the count of multi-family units.
Ruan’s post-graduation plans include growing the real estate startup company Orbbit. Working as the company’s technical co-founder, she is building a LinkedIn + JIRA program powered by AI.
Ruan’s research was completed under the guidance of Eric Chi, associate professor of statistics; Katherine Ensor, the Noah G. Harding Professor of Statistics and director of CoFES; and Elizabeth McGuffey, associate teaching professor of statistics.
- Shawn Hutchins, Communications and Marketing Specialist
