Abstract

In this TREO talk, we discuss research approaches and tools to investigate the relationship between innovation and location variables associated with universities. The research team has data on different indicators of innovation, such as filed and issued patents, the number of startups formed and closed in the area, license incomes, research expenditures, and demographics from more than two hundred universities. The data has been reported by universities for more than two decades. Complementary data, such as institutions’ endowments, the presence of engineering schools, prevalent occupations, crime index, and weather-related characteristics, were collected to enrich the original dataset, which now includes more than 60 variables. This dataset presents a valuable opportunity to analyze how universities’ efforts, demographic factors, and location-based characteristics influence different innovation outcomes. Prior exploration of a five-year subset of this dataset suggests that variables such as income per capita, median house value, student-to-faculty ratio, LGBT friendliness, foreign student population, and walkability correlate with a measure of university investment efficiency: Technology Transfer Income/University Research Expenditures (CARE, 2017). Motivated by recent CAIS research, we focus on location-related data, some of which is already available, while other variables need to be collected. According to "Location Analytics in Information Systems: Opportunities for Research and Teaching" (Erskine et al., 2024), one way to advance research in this area is to revisit prior studies, incorporate spatial data, and explore how Location Analytics (LA) can enhance findings. We embrace this approach through the lens of university-driven innovation. Several challenges arise in this analysis. Universities vary widely in structure and business models, and self-reported data may refer to either a single institution or a university system. Some universities have a strong research orientation, while others focus primarily on teaching, with many falling in between. Additionally, innovation is a dynamic phenomenon influenced by multiple factors, often supported by anecdotal rather than empirical evidence. During our presentation, we will outline our approach to addressing these challenges and the tools we are using to clean and enrich our data, including generative AI. Following Erskine et al.'s recommendation, we are exploring advanced LA techniques such as geo-segmentation and spatial autocorrelation. We look forward to an engaging discussion on these and other analytical techniques, as well as the broader theoretical considerations for our study. We aim to foster meaningful dialogue and collaborations that will help refine this research as it evolves.

Comments

tpp1035

Download

Share

COinS