Data Wrangling Workshop Post-mortem

23 February 2018

In the Data Wrangling workshop, I tried something a little different: I interleaved portions of the DataCamp lesson on dplyr with exercises of my own design that had learners redo their earlier work from the Spreadsheets lesson using the vitals dataset, in R.

The DataCamp lesson had short videos explaining a dplyr command, followed by a few basic exercises. Learners then had the chance to apply the same concept to the vitals dataset in RStudio, with a little less hand-holding.

I thought this session worked really well. Learners seemed to like the DataCamp videos (we watched the introduction together as a group), and liked the dplyr framework as a whole.

I saw some eyes light up when we got to the section on pipes. With about 30 minutes to go, I switched teaching interactively. I quickly summarized the commands up to arrange() and then explained piping as an alternative way to string together dplyr commands. I also then demod table joins aswell, because I really wanted everyone to at least see that it was possible, and very easy.

Takeaways:

  • Learners enjoyed the DataCamp video/drill + working with a dataset approach.
  • 2 hours in not enough time. As per always.
  • I’d like to try more teaching interactively. Learners seem quite comfortable following online lessons, but I might try doing a quick (30 min) interactive lesson that covers a chunk of the important content (in this case, ggplot is up next) interactively, and then set the learners loose to work on their own.
  • … Well, not on their own, I still encourage people to work in pairs. I could probably do a better job of encourage that. People seem to need push.