research overview
- My research interests are rooted in geospatial data science, extended to both natural and social sciences. I contribute to and use methods from the fields of spatiotemporal machine learning, spatial statistics, human-centered visual analytics, and remote sensing. I view my research as an integrative activity that brings data science together with environmental and social sciences to help make society more sustainable and equitable. My research program is use-inspired, and therefore I hope it to be high-impact, well-funded, interdisciplinary, collaborative, and stakeholder-involved for advancing science in the context of pressing environmental and social issues. At the core of my current research, my work is focused on harnessing non-linear spatial and temporal relations for data-driven and human-centered computing. Concretely, this line of inquiry has taken the form of designing and evaluating spatiotemporal machine learning and quantitative methods for applications in various domains, including environmental remote sensing, epidemiological modeling, public health, social media analytics, and renewable energy systems. My research is computational, and it involves wrangling large heterogenous datasets as well as designing, programming and experimenting with algorithmic models and/or visual interfaces on advanced hardware. Additionally, most of my projects involve stakeholder engagement to ensure relevancy, adoption, and impactful science. Developing methods and applications in the context of real-world problems, where there are experts in other fields, will mean that the methods need to be useful and usable. We develop methods in the context of real-world applications, meaning most of the time, we are not using off-the-shelf clean benchmark datasets. While is challenging at times, doing use inspired research has helped us stay rooted in reality, guide our assumptions and decisions, establish a funded research program, and hopefully, have more immediate impact.