Modelling Count Data in R: A Multilevel Framework

A Quick Practical Guide

One of my research areas of interest revolves around understanding human mobility and migration flows. Key attributes of flow data are that they are counts and a right-skewed, overdispesed and often zero-inflated distribution. And I was interested in measuring the variability in region-specific intercepts and slopes in a multilevel modelling framework in R! So, I decided to test the most commonly used R packages to fit four different variants: Poisson, Zero-inflated Poisson, Negative Binomial and Zero-inflated Negative Binomial models - and I wrote a quick practical guide to do this using three R packages: glmmTMB(), glmer() or glmer.nb(), and lme4. I wrote the tweet thread when I did this here:

If you are interested in the guide, click HERE

Francisco Rowe
Francisco Rowe
Professor of Population Data Science

My research interests include human mobility and migration; economic geography and spatial inequality; geographic data science.