I have two main research themes: causal knowledge analytics and health information systems. I use diverse methods in my research, including graph theory, network analysis, natural language processing, large language models, and experiments. I am passionate about working on the following two research themes and my methodology and research training also have prepared me to explore other topics.

Causal Knowledge Analytics 

Advancing scholarly productivity with digitization and analytics

The development of scientific fields produces a large and exponentially expanding knowledge base, which greatly limits scholarly productivity. My dissertation aims to develop methods to assist scholars in processing the literature by digitizing and analyzing causal models. The foundational work of my dissertation is to digitize knowledge as graphs and store them in a property graph database. To build a database for my research, I have coded about 300 publications in 13 MISQ curations. The resulting graph database opens opportunities to assist literature learning using graph-related analytic methods. We define this line of research as Causal Knowledge Analytics:

Causal knowledge analytics is concerned with developing and applying methods for processing graphical representations of causal models to advance theoretical research by scholarly communities.

My dissertation builds a five-level causal knowledge analytics research framework, which develops a set of reproducible analytics to defragment knowledge and improve knowledge accessibility, synthesize knowledge, and facilitate longitudinal and field-level examinations of IS knowledge advancement. To support the proposed analyses, this dissertation uses methods including graph query language, social network analysis, natural language processing (NLP), and graph theory. 

Mobile health

Promoting health behaviors with IS design

Since the design of mobile health apps can impact people’s health behaviors, it is important to study and seek effective designs. A distinctive aspect of technology is the fact that, unlike the offline world, it allows for a more fluid construction of reference groups. Yet, there is little empirical evidence of the efficacy of different social comparison groups on exercise behavior, and traditional approaches, such as generic reference groups, may not necessarily be the most effective. We aim to identify more meaningful ways of constructing reference groups based on social comparison theory.