search

r4wrds / r4wrds

Water Resources Data Science R course website
https://www.r4wrds.com
31 stars 7 forks source link

Containerization and code parameterization for reproducibility #80

Open ksonda opened 3 years ago

ksonda commented 3 years ago

Wondering if you would consider adding material around using docker to containerize particular R and package versions, and create reproducible workflows by combining them with parameterized scripts that allow you to rerun analyses from CLI.

richpauloo commented 3 years ago

It's a good idea Kyle, and one I've spoken with Ryan and others about.

I think CI/CD topics may fit best into an "Advanced" course, or maybe a "productionizing" course. For most R users in our field, I would wager that containerization, building packages, writing tests, deploying apps and programs to the cloud, and working from the CL - although very important - are less useful on a day-to-day level than the skills that advance our ability to do science, like functional programming, interactive visualization, EDA and so on. However I agree that the skills are very useful for a smaller group of R users, and it's my goal to find funding to develop those advanced materials at a later date. Lmk if you want to help with that in some way.

ksonda commented 3 years ago

Ok, if you get around to it I can probably contribute some stuff in this general area. https://github.com/dblodgett-usgs/hydrogeoenv might offer a good little starting place.