What I use

Here is a list of tools I use on a daily basis. Feel free to get inspired!

Coding

My preferred language for data wrangling is R and the {tidyverse} family of packages.
I am partial to Python, especially since the release of multiple R-like libraries, such as polars, shiny (equivalent to {shiny}), great tables (equivalent to {gt}), plotnine (equivalent to {ggplot2}) and Shinylive for Python (equivalent to Shinylive for R).
Reasons to use Julia are Dynare for Julia and Tidier, a reimplementation of the {tidyverse} family of packages in Julia!
I use plotly to create interactive dashboards.

When it comes to creating visualizations, I try to follow the book Storytelling with Data by Cole Nussbaumer Knaflic.
I try to follow Hadley Wickham’s book Tidy design principles, which builds on the Tidyverse style guide, which in turn is derived from Google’s R Style Guide. I use the {styler} R package and RStudio add-in to re-style existing code with a single click.

I am fluent in LaTeX but I would recommend typst for typesetting beginners.
I have completely moved from R Markdown to Quarto, but I still recommend the R Markdown Cookbook to learn more about Yihui Xie’s amazing {knitr} package.
I read and annotate PDFs with Zotero.
To make PDFs more visible on GitHub, I use nbviewer.
When working with Overleaf, I use Better BibTeX for Zotero to automatically generate citation keys without key clashes.
I highly recommend the the Monash Quarto Templates and the Quarto Journal Templates.
I use The Rogue Scholar to create stable DOIs for each of my blog posts.

I use the {renv} R package to snapshot the R version and the package versions and save them in an isolated library.
When I want to snapshot the entire ecosystem, including system-level dependencies, I use the {rix} package.
I use {targets} to create Make-like pipelines.
I try to follow Bruno Rodrigues book Building reproducible analytical pipelines with R and Building reproducible analytical pipelines with Python, respectively.

I have been experimenting with GitHub Copilot in RStudio.
To create GitHub Actions workflows, I usually ask ChatGPT.
If you want to properly cite your sources, I recommend perplexity.ai.
The Claude chat bot is useful when developing large projects.
I use {ellmer} to interact with LLMs from my IDE.
Check out the BIS Quarterly Review Special Feature Large language models: a primer for economists.