I've been reflecting on a peer-reviewed abstract I co-authored and presented at the 2022 DTS (Diabetes Technology Society) conference, entitled Diabetes Phenotypes Quantified via a Physiofunctional Model Fitted to Raw CGM Time Series Data. Here linked is the final author's manuscript copy of the published abstract.
This was some of the first quantitative modeling I had done at such a large scale--- tens of thousands of individuals represented in a multi-million row dataset, comprising many years'-worth of diabetes-focused health monitoring efforts.
I collaborated daily with colleagues who diligently applied themselves to curating well-labeled samples, ensuring a well-defined hypothesis space was not only possible but was guaranteed via sound statistical reasoning.
This experience really opened my eyes to a few key truths, namely the degree to which Big Data can capture actionable insights with remarkable quantitative certainty.
The possibility of such profound real-world impact has further motivated me to be meticulous in my work.
As founder of PFun Digital Health, I'm developing an evidence-based inference method that enables automated generation of personalized health tips (click to try the live PFun Health Tips web demo).
Persistence has been the only way forward; I'm currently still working on this solo, without any formal organizational affiliation. There are days I question if my intent is to experience pure sysiphean punishment. After reflection though, I can state with confidence that my trudging progress is borne of a desire to continue the momentum of that first impulse, to finally reach out and touch someone.
