I'm not sure exactly where to post this, but this forum seems like a good place to start.
Frequency tables need some Tidy Love
Tidy methods for quantitative data & models have advanced considerably, but there hasn't been much
development of similar ideas for "categorical data", by which I mean data that is often compactly represented
as n-way frequency tables, cross classified by one or more discrete factors.
I've written a draft vignette for the vcdExtra package, Tidy Methods For Categorical Data Analysis where I discuss this topic. I outline existing methods for creating and manipulating such frequency tables, fitting (loglinear) models and graphical methods, and suggest some ways that this could be brought into a tidyverse framework.
I'm hoping for some interest from tidyverse developers in making this long-time dream a reality,