inlaops - an introduction

This is a project that I wanted to have working for a while.

Working in the life sciences sector I learned that INLA is a very popular package for modeling.

However, making it fit in the wider R ecosystem is a challenge. There are a few gaps, but in particular I was looking for a way to have a deployed model.

I thought that using it with vetiver should be the way to go, and so inlaops came into being. My specific case involves hosting on Posit Connect, but an API can be published in many ways, and this makes it suitable for different environments.

With inlaops one can follow the usual steps of fitting the model, and then making the model available on an API.

The Posit Connect use case is documented on the package’s site.

inlaops was made possible first by being able to install INLA from a local package manager, and then from inlabru having introduced Tidy Postprocessing.

I expect there are some rough edges in the package. It works on my demo, but I would be interested in hearing feedback from users.

The initial release is on its way to CRAN.