Personal Highlights: Positconf 2023

Data science

Curation of content to check out when I’ve got time

Author

Chi Zhang

Published

September 21, 2023

The yearly party of Positconf (formerly Rstudio conf) has come to an end. I joined the virtual experience at home, it is of course not the same as attending in-person, yet the atmosphere in discord was still great!

It’s hard to choose which talks to watch since multiple were scheduled at the same time, so one has to prioritize. I definitely will re-visit some of the talks at a later point, so this blog acts as a placeholder for links so that I can find them in the future.

Make interactive things

WebDev is definitely a big thing at this year’s positconf. If I’m learning one thing from the conference, I’d check out webR.

I still remember when R was mainly for statistical analysis and computing back when I learned it. Now it’s become much more fun! Strictly speaking, webRand quarto are not R per se. However, they’ve become the gateway drugs for R programmers to dabble in WebDev. With web assembly (wasm), now one can execute R code in a browser and even run shiny app.

Unlock the power of dataViz animation and interactivity in quarto by Deepsha Menghani used a super fun example (F-bomb) to demonstrate how to add interactivity to your barplot (or other plots) with Crosstalk. Check out the talk here. The presentation was as interactive as the quarto slides, good job Deepsha!

Running shiny without a server by Joe Cheng (repo): this was a big announcement. I used shiny at work, but for my own projects or smaller teaching projects I tried to stay away from shiny - I was concerned about the fee. This looks like a promising thing to try out once it’s stable, although I’d probably do webR first.

Make pretty things

It is fascinating to see so many organizations and individual R developers make their own themes for better branding, recognition and storytelling. More and more peple have realized that making beautiful plots is important, and totally possible as well. Work on layout, color, font and sizes!

Themes

Adding a touch of glitr: Developing a package of themes on top of ggplot by Aaron Chafetz, Karishma Srikanth and colleagues at USAID. repo

Tables

Making tables with gt has been on my to-do list for a while now. It is very inspiring to see so many cool tables that makes you wonder, “is it really JUST a table?” For example, check out this gallery by Posit community.

The book Creating beautiful tables in R with gt by Albert Rapp would be a good place to learn how to make nice tables. Actually the reason why I wanted to use gt is that it seems to be the mainsteam in clinical reporting in pharma. I bumped into this blog post some time ago, and this would be my starting point.

Quarto

If you want to go one step further and start making your quarto project pretty, there are a few things to try out.

Albert Rapp in his talk HTML and CSS for R Users stated that quarto is a gateway drug to WebDev. It reminds me of my very first presentation at my local R users community (2019) was about building a website with blogdown, and when I really spent a lot of time to make my markdown documentation colorful with span style - and that was about everything I knew.

Now I want more. Learning HTML and CSS can make your dataviz, tables, slides and dashboards look not only professional but also special. I’m going to check out the scss variables in quarto which defines the theme, theme_file.scss. Emil Hvitfeldt (Styling and templating quarto documents) showed us how to make really pretty and animated (!) quarto sldies themes, and shared this template with us, quarto-revealjs-earth. I really like how revealjs slides look like, just that the MacOS Keynote (or MS ppt) drag-and-drop seems more flexible to me (?) Guess it’s something I should get used to over time.

Richard Iannone (Extending quarto) introduced quarto shortcode extensions to add a bunch of fancy-looking icons to quarto files. To create extensions in general: https://quarto.org/docs/extensions/creating. This is for more pro-users since you needs to learn lua.

Quarto updates

Quarto is definitely one of the most discussed topics in the year 2022-2023 in the R community. For good reasons. I need to catch up the the latest developments annd use-cases:

and so many more. I couldn’t follow all the talks and I’m sure there are lots of great examples of how quarto is better than traditional ways of reporting.

A few other things to check out

Beyond the web and quarto topics, I think there are some existing and new tools that can be useful for my work. For example,

  • I should review Hadley and Jenny’s R package book (2e).
  • this package targets for pipeline automation and management look like something that can be used for my analysis
  • …

It will take a while to digest the latest developments. But little by little, we’ll get there! People in the R community are doing great things.