Projects
As of August 2020, I am a member of Dr. Peter Song’s research group where my research focuses on information geometry-based causal discovery and inference. I have recently started exploring conformal predictions as well.
I’m also a part of the NIH-funded multi-centre Diabetes Foot Consortium of clinicians and analysts where we study diabetic foot ulcers, a common complication of diabetes and the leading cause of lower limb amputations in the United States. Led by Dr. Cathie Spino, my work involves the statistical analyses of ulcer-related wound images.
Since March 2020, I have also been working on studying and modeling the spread of infectious diseases with a group of statisticians, epidemiologists and economists led by Dr. Bhramar Mukherjee.
Conformal predictions (2023-)
Click here for more details
This part is still under construction! Please check this place again soon :)
Information geometry and causality (2021-23)
Click here for more details
Investigating public health disparities requires identifying underlying causal factors. This project forms the main focus of my dissertation. I propose novel and scalable analytic methods for causal discovery and inference with special emphasis on public health. Said methods bridge statistical and information theory for the sake of application to biomedical research. To solicit the interest of practitioners from diverse backgrounds, the proposed methods are accompanied by well-documented software.
Diabetes Foot Ulcer Consortium (2021-23)
Click here for more details
The Diabetes Foot Consortium is a multi-centre consortium funded by the NIH. In addition to performing standard statistical data analyses for the consortium, I have built a web-based dashboard that allows clinicians to access and download graphical and summary statistics associated with the study. In this dashboard, I was able to automate some data pooling and analysis pipelines, allowing for faster sharing of findings within the group. In addition to my work as an analyst, I took part in writing grants for ancillary projects within the consortium.
Oceanography and machine learning (2020-23)
Click here for more details
One of my earliest collaborations was with a group of researchers from India who study trends in fish harvesting and its impact on aquatic life in the eastern part of India. I wrote software for implementing a Bayesian time series model that was successful in detecting change points in fish catch data in Eastern India. Since a lot of coastal communities in India rely heavily on fishing as the primary source of income, being able to predict fish catch is critical for advising policymakers on labour allocation as well as providing additional support during times of reduced fish catch. This is, in turn, helpful for improving the lives of coastal communities in India. My collaborators have helped me not only with analyzing oceanography data but also with understanding the massive impact of climate change on countless fishing villages along the coast of India. I have continued with this collaboration and am actively involved with a new machine learning-based project for time series forecasting of fish catch data. For more information on this project and related papers please see my publications page.
The COV-IND-19 group (2020-22)
Click here for more details
Ever since COVID-19 became part of our lives, I have been involved with a group of researchers from the University of Michigan, Johns Hopkins University, Harvard University, Imperial College London, and other institutions from India, the UK, and the US to study and model infectious diseases. Specifically, I have been interested in the spatio-temporal forecasting of infectious diseases. Among many other challenges in improving health disparities, COVID-19 was a stark reminder of how underprepared humanity is, in terms of informing and protecting those with an elevated risk of illnesses during a pandemic. I hope that through collaborative research, we have an improved understanding of how diseases spread and have differential impacts on many communities. For more information on this project and related papers please see my publications page.