Prof. Francesca Dominici, the Clarence James Gamble Professor of Biostatistics, Population, and Data Science at the Harvard T.H. Chan School of Public Health and Faculty Director of the Harvard Data Science Initiative, will join the “Understanding Climate Change in South Asia” panel at LMSAI’s Annual Cambridge Symposium: Science and Technology on Friday, May 3. As a data scientist, Prof. Dominici develops statistical methods and machine learning approaches to look for patterns that address complex public health issues, such as air pollution, noise, and climate change. We spoke to her about her research in the Q&A below.
Mittal Institute: Prof. Dominici, thank you for speaking with us. What initially drew you to the field of environmental biostatistics?
Francesca Dominici: I’ve always been very passionate about mathematics and probability. I was initially drawn into statistics, and then I also wanted to combine this passion with social good—to do something that helps the world.
For my first job, I was hired as a postdoctoral fellow to develop statistical models to understand the health impact of environmental contaminants. That’s when I realized this is an area where more research is needed.
So, for me, the combination of a passion for mathematics, statistics, and the prospect of doing something good for society in terms of addressing air contaminants became my lifelong career in understanding the health impacts of air pollution and climate change-related stressors.
Mittal Institute: Can you describe your research – how can Big Data shed light on the health impacts of climate change, and how can it impact policy measures?
Francesca Dominici: I think it’s really important to recognize that my paradigm and the framework that drives me is to use science to inform policy. I’ve always aimed at producing the most rigorous, scientific evidence to address policy related questions. And when you’re doing science in the context of the health impacts of environmental contaminants, or the health impacts of climate change related exposures, data is essential. There is no data science without data. I’ve been devoting a tremendous amount of time and resources to go after the best possible data on health outcomes, and everything we think of when it comes to climate stressors related to exposure to wildfires, exposure to air pollution, exposure to heat, etc. I take all of this data from government data sources, and we call it multi-modal because the data is coming from images, electronic medical records – it is all very heterogeneous.
We develop what we call a multi-modal AI approach to integrate all of these data into a common platform, and then we develop causal informed AI models. We then ask whether or not exposure to air pollution causes health outcomes. Developing these new AI algorithms allow us to interrogate the data with questions; for example, if we were to reduce air pollution by even a small amount figure, how many lives would we save?
Mittal Institute: What are some things you hope the audience takes away from your climate change panel at our upcoming symposium?
Francesca Dominici: I hope they go away with an “infectious optimism.” I think so much about the climate change narrative is that “it’s going to be terrible,” or ‘we have not met any goals,” which, unfortunately, is mostly all true. But I hope to instill a sense of optimism in terms of our development of data tools; we are developing new technologies that hopefully will allow us to respond very effectively to this climate crisis. We need to come together to do this. And I also think that thanks to the fact that we are all better connected than we used to be – that we can accelerate knowledge and training of people doing data science in other countries – we can analyze our data more effectively.
The views represented herein are those of the interview subjects and do not necessarily reflect the views of LMSAI, its staff, or its steering committee.
