Point Process

Many distinct mathematical models represent where and when "things happen" and Point Process formalisms make it possible to combine these models under one inferrential umbrella. A single estimation stage can combine Poisson, exponential, gamma, and various other models with spatio-temporally varying hazard rates. The price is some mild math.

details

Using Small Area Estimates

Using counts from small areas as denominators (or numerators) in producing statistical estimates can be difficult as denominator uncertainty can exacerbate uncertainty about the overall rates and standard methods often produce non-sensical (e.g.-negative) count estimates when uncertainty is considered. Tract-level population denominators show some of these difficulties but they can also be addressed comprehensively to enable more detailed spatial studies.

details

Mapping the Social Vulnerability Index

The Social Vulnerability Index (SVI) ...

details

Contact

I am a statistical epidemiologist with experience modeling events based on surveillance sources and medical trials. My research focuses on creating decision support tools for public health challenges by using mathematical modeling, Bayesian inference, and direct collaboration with decision-makers. Past projects addressed dengue fever control, spatial models of contraceptive access, and the comparative evaluation of new treatments in oncology. I am currently a Senior Research Area Specialist in the Department of Epidemiology and the Center for Social Epidemiology and Population Health at the University of Michigan.

krzysztof@mostek.io