Highlights from Cascadia R Conference, Drafting a Route Profile Dashboard, Hex-bin Statistic Mapping
This content was presented to Nelson\Nygaard Staff at a GIS Hangout on Wednesday, June 9th, 2021, and is available as a recording here and embedded below.
I (Bryan Blanc) attended the main, non-workshop segment of the 2021 Cascadia R Conference on Saturday, June 5th, 2021. Esther Needham and I presented regarding our work on the King County Metro/SDOT iteration of the Bus Delay Analysis Tool (BDAT). There were a number of talks that introduced me to packages I was unaware of, as well as educational resources, and I wanted to highlight those for NNers below. I am working on getting recordings of talks (including our own) I found particularly germane to our work and resources I link to below. I will update this page as I get ahold of recordings.
Here were all of the sessions of the main conference day:
A number of the talks focused on R education, both within the context of an agency/firm as well as a more traditional (university/college) educational context.
Office Hours. One idea I thought would be good to implement immediately at NN is for me to have ‘office hours’ – i.e. a consistent time where anyone can drop in and ask questions live without having to make an appointment. The person who suggested it said that the easy/drop-in nature helped to facilitate learning, and to continue to make it remote in perpetuity especially given our multi-office nature. I am going to propose I do office hours every other Wednesday – i.e. the off-week in which we do not have a programmed R training scheduled, although I could do them more often if people think they would be helpful, or if we find that the one-hour, biweekly slot is consistently full of questions/activity.
R for the rest of us. The creator of ‘R for the Rest of Us’, David Keyes, was presenting at the conference on a package I will discuss below, but I wanted to highlight his website as an educational resource. The first course is free, but the other ones are paid. I haven’t looked deeply enough into this to see how it compares with DataCamp, but if anyone wants to try out the free course and report back to the group of its usefulness, we can encourage folks to consider this as an alternative to DataCamp – the paid courses could be paid for with professional development dollars, and we could look to see if there is any benefit to people taking the course as a group.
DataCamp. Many of you are probably already aware of DataCamp, but I wanted to point it out for any of our newer folks. It’s a great way to learn R skills interactively on the web. I have an account and still use it to learn new things. You can try out the free courses, and there is a paid subscription for the bulk of the content, which you could pay for with professional development and/or tech funds. They also just added a new feature in which you can form a group/organization in DataCamp. You can join the NN organization with this link.
TidyX Screencasts. One of the main organizers of the conference, Ellis Hughes, hosts a weekly screencast on R topics on Youtube. I haven’t watched any of the episodes yet but they look interesting!
Tidy Tuesday podcast. The above also made me stumble upon an R podcast 🤓😆, Tidy Tuesday ! For folks looking for additional podcasts, maybe consider this one.
Ready for R. Another free course put together by one of the conference organizers, Ted Lederas, is Ready for R, which was assembled for an OHSU (Oregon Health and Science University) course. Again, I have not checked it out myself, but encourage other folks to check it out and report back to the group.
There were a bunch of packages I didn’t know about that either presenters at the conference had created themselves or they mentioned in the course of their presentation. I have tried to catalogue all the ones I noted below for further investigation.
pagedreport looks like an impressive/easier way to generate PDF reports from R markdown. There is a good demo of it in this blog post.pagedown, which pagedreport extends, is an RStudio supported package to paginate HTML based markdown content into PDFs more easily, so if you had already set up markdown content for a web page (like this blog), you can paginate it for a static PDF more easily with pagedown.NMFSReports looked like a great package for generating technical memos, reports, and presentations. It was developed and demoed by a scientist for NOAA, but I imagine could be fairly easily extended/customized for other styles, e.g., Nelson\Nygaard.golem looks like a great package to turn a Shiny app into… a package! The idea is to maximize reproducibility for production-grade applications. I would like to explore this for both the Route Profile dashboard applications as well as BDAT.renv looks like an improved version of packrat, which will preserve package versions to production-grade code. Esther and I recently discovered an error in BDAT that stemmed from us each using different versions of the same package. While the application is still in production, this is fine, but when putting ‘pens down’ on the application, it would be nice to know that future package changes will not break the application. I would like to experiment with this for BDAT.fuzzyjoin is a package to join tables on inexact character matches. This relies on the stringdist package, which I have used before to do this manually, but the approximate string matching is integrated into the join functions directly. Sometimes you will get inexact character string matches. I have encountered this when trying to match stop level data to GTFS where the client only has stop names, which inevitably do not match precisely to the GTFS character strings of the stop names.patchwork looks like an easier implementation of grid packages for arranging multiple ggplots on a single plotting canvass.Paul will share the draft StarMetro route profiles webpage he’s developing for the StarMetro COA. He will share details on adding icons to a leaflet map, and other details about reactivity of the leaflet maps. Code is saved here: G:\Current\STARMETRO_2021_Update_2020.0099\R\star_metro_route_profiles
Ulises will share the suitability analysis implemented to identify potential locations for the expansion of the Cap Metro Bikeshare system in Austin. He will share details on how to create a hexagonal grid and intersect different feature types (lines, points and polygons) to create individual indicators. Code saved here: (work in progress, so might not be the easiest to read). G:\Current\CAP METRO_MetroBike_Strategic_Exp_2020_0955\R\cap_metro_bike_expansion_atlas_maps_20210420.R
This content was presented to Nelson\Nygaard Staff at a Lunch and Learn webinar on Wednesday, June 9th, and is available as a recording here and embedded at the top of the page.