Tips & Tricks from the newbies of rstudio::conf2020
To support the professional development and continued learning of our team, the ACLU analytics team encourages people to attend at least one conference of our choice per year. In 2019, our very own Brooke Madubuonwu spoke at the R Studio conference, and inspired Nikki and Sophie to attend this year! Here are a few of their favorite presentations, most immediately-usable learnings and tips and tricks for any newbies of rstudio::con2021.
Nikki:
Nikki is a data analyst on the Analytics team at the ACLU. She analyzes data and builds dashboards using our BI tool and uses R to support both the fundraising and legal departments. Nikki loves the janitor package and couldn’t live without clean_names().
Sophie:
Sophie is a data journalist on the Analytics team at the ACLU. She spends half her time analyzing data (using R!) to help lawyers with their active cases and the other half of her time making data visualizations and graphics for public-facing advocacy pieces. Though (like Nikki) Sophie loves nothing more than a clean dataset, her favorite R package is obviously ggplot2.
1. Presentations and R packages we found useful and relevant to our work?
Sophie: I enjoyed a lot of the R markdown and data visualization-related talks. A few, I’d even say, changed my whole life.
- This presentation by Will Chase about the “glamour of graphics” was an incredible overview of the “rules” for what makes a visualization easy to read and appealing to look at. Some takeaways were to 1. download a color picker that allows you to create your own customizable palette and 2. lots of whitespace and the right typography can go a long way in making your charts look clean.
- This presentation by Desirée De Leon went behind the scenes on her brilliantly-crafted stats lesson using hand-drawn “teacup giraffes”! Not only is what she built an approachable resource for anyone that wants to learn R and basic statistics, but the code she used to build it is available on Github, and was made entirely using R Markdown. Desirée really inspired me to help our team revamp our R Markdown template design; we may present our reports mainly to lawyers, but that doesn’t mean they can’t be pretty.
- Nick Tierney’s presentation on “making better spaghetti (plots)” was all too relatable as someone who often needs to make sense of many data points and freaks out over the use of > 3 colors in a plot. In his talk, Nick introduces the “brolgar” package, made specifically for identifying patterns in messy longitudinal data, and shares some exact steps to take the next time you find yourself in a spaghetti viz nightmare.
- Everyone that uses ggplot2 should know about Claus Wilke’s package “ggtext.” The ability to easily control the look and feel of text in your graphic was well overdue — I immediately made use of this package upon my return.
Nikki:
- A lot of my work requires the use of our data warehouse, which is a Redshift server. I learned SQL before learning R and found the two languages had wildly different learning curves. SQL’s syntax is much more intuitive to me than R, though learning dplyr has helped really helped bridge the gap between SQL and R for me. Whether you are just starting to learn R, or need queries translated from SQL to dplyr, the package “queryparser” can help. Before you get too excited, “queryparser” has its limitations that can hamper many of your query translations.
- Object of type ‘closure’ is not subsettable was given as a keynote address to the entire conference. I found this talk particularly useful because we all deal with indecipherable error messages while using R and sometimes we don’t know where to begin. Luckily, Jenny Bryan provided a useful framework on how to approach these errors. Her suggestions range from restarting your session to more complex ideas, such as creating a reprex for others to debug, have already been useful to me when troubleshooting my code.
- I very much liked Sharla Gelfand’s talk titled “Don’t Repeat Yourself, Talk To Yourself!” I work on a few projects where we receive new datasets every few months that require an update of the previous analysis. In order to be as efficient as possible, it’s important to make code reproducible for future use. A person can do this by: 1) reorganizing your files to make things easy to find 2) make functions 3) remembering that artisanal data does not require artisanal code and 4) automating templates. If you, too, find yourself rewriting the same code for similar projects repeatedly or feeling unable to remember which _FINAL_for_real.R file is really the right script to use, watch this talk and immediately reorganize your projects — your future self will thank you.
2. Favorite presentations completely unrelated to our work?
Sophie: When you love something enough, make a (pocketed) dress out of it.
Nikki: One of the most fascinating talks was about Legos (!). Yes, Legos. Ryan Timpe spoke about how he built the package “legoR” as a side project. This package lets you upload an image of anything (your pet, a selfie… etc) and then pixelates that image and provides you with the exact legos necessary for you to build that thing IRL.
What started as a project in Ryan’s free time, ultimately landed them a job at LEGO as a Senior Data Scientist.
3. Things we’d recommend for first-time rstudio::con folks:
We (Nikki and Sophie) highly recommend rstudio::conf2021 to anyone that uses or is interested in R. While there was an overwhelming amount of interesting information, many presentations were immediately useful and the people we met were approachable and came from all sorts of R backgrounds. A few things we’d recommend for newcomers (things we wish we knew or did better!) are:
- Download the app and really plan out your schedule! Both of us created a rough outline of our daily schedule, but once the day got going, things moved really quickly and it became hard to look up which talks seemed most interesting in between presentations. If you go with a coworker, split up if you find two talks equally interesting so you can report back to each other. If you really can’t decide, “pull a Sophie” and stand directly outside the hallway of two presentations and walk back and forth between the two until you decide.
- Join the “Birds of a Feather” (BOF) groups during lunch and break times. BOF set up lunch tables for people with similar interests to meet each other. We didn’t fully understand their purpose until day two, but had a great time meeting people in our field and wish we joined in earlier. Related: try and find your favorite speakers, and get to know them a little!
- Get. Those. Stickers. Stickers at the rstudio conference are extremely popular. If you see a bunch of folks crowded around a table, there are probably stickers on that table, and they will go quickly.
- Book your conference tickets early. There are a limited number of tickets available at a discounted price, so be sure to snag them as soon as you can, especially if you work at a nonprofit. (Note: looks like early bird tickets for rstudio::conf 2021 are already sold out!)
