Bits and pieces

This entry collects some quick links and tips.

🔗 a good distribution for count data (DOI: 10.1111/2041-210X.14387)

📌 use parallelly::availableCores() instead of parallel::detectCores() because of problems with the latter

🔗 flint R package for fast code linting and fixing lints. I like how the author tested the package by linting and repairing huge R packages (e.g., ggplot2 or targets) and getting the patches accepted.

Default for an NA value. A trick in the linked post on parallel::detectCores() to replace a possible NA with a safe default: max(default, possible_na, na.rm = TRUE).

🔗 https://0.30000000000000004.com/ - I can never remember how to search for this URL.

📌 Two-tailed p-value: 2 * pt(-abs(t), df), or to stay on the log-scale log(2) + pt(-abs(t), df, log.p = TRUE).