#1

Hi there!
Is there a specific rule on when a new category gets created? For instance, if there was interest in creating a GIS/geospatial category of topics, what would need to happen in order to make it true?

Sorry if it was covered in F.A.Q. somewhere - my Google-fu wasn't effective enough to find it.

#2

I'm under the impression that since this forum is sponsored by RStudio, only topics directly addressed by the tidyverse and RStudio ecosystem deserve a separate category and everything else should go into the #general category with proper tags. Although, I would vote to include sub categories into #general to better classify broad topics.

1 Like
#3

I would love to see a spatial category! There are the #ml and #package-development categories, which are not R Studio/tidyverse specific...

1 Like
#4

For #ml there is tymodels and friends and #package-development is addressed by the IDE and other rstudio tools. I didn't mean exclusively addressed by RStudio.

1 Like
#5

@andresrcs hit the main points of why a spatial category doesn't exist yet. There was also thought that spatial should just be a tag until demand for something else arose.

We are also concerned about having too many categories, as categories suggest silos, while many issues cross multiple categories.
For example, I suggested the #ML category have two sub-categories, one for Tidymodels and one for MLverse topics (eg our AI blog, sparkyr, pins, and other distributed computing topics), but those folks felt we should instead be supporting these projects with a set of package tags, and keep the modeling general modeling category.

I'm very much open to a spatial category, I've been reaching out to other folks I see active in the spatial data science community for their advice.

Would you mind listing the spatial r-packages you like to use the most? ANd other resources that you depend on in your R+mapping work?

2 Likes
#6

Sounds great! In my work, the most essential spatial package is sf. It plays very nicely with dplyr/tidyverse functions and has methods for all the main dplyr functions.

Beyond handling and manipulating spatial data, some key packages for making static maps are ggplot2 (with the somehwat recent geom_sf() and tmap. For interactive maps, there is leaflet (an RStudio package!), mapview, plotly and others.

There is a more exhaustive list here: https://www.r-spatial.org/projects/ which links to a couple of CRAN spatial task views as well as including GitHub only packages.

Geocomputation with R is a great book that is available for free online. And #rspatial is pretty active on Twitter. Finally, @angela recently started a weekly R spatial chat on Twitter

2 Likes
#7

The spatial task view folks currently suggests the following for people who'd like to pose questions, and help out.

I'd like to politely make the case that a list-server is not an ideal place to direct the types of discussions likely to occur here. A major motivation to introduce RStudio community was that non-developers (that is, statistical practitioners, learners, and even many super-users), were not having a good experience with the relatively unforgiving culture on stack-overflow and R-help list-servs.
To be clear, I am unaware of this issue with the R-SIG-Geo mailing-list.

Further, open-source software development tools, like github and listservs, just aren't optimized to act as both the primary location for software development and to help learners deal with novice questions. They also don't scale well as many novice users seek help, lack UI features that many moderators deem essential, and many are not as optimized for SEO.

1 Like
#8

Personally, I've never signed up for R-SIG-Geo for many of the reasons you mention, @EconomiCurtis. I also have no idea how beginner-friendly it is, but the Discourse format is so much more conducive to actually getting help. I'm not sure the best way to demonstrate this other than seeing the number of questions that get answered in a quick and friendly way here! @jlacko is one of the community members that answers a lot of spatial questions.

2 Likes
#9

Thanks for the pointer for an interesting conversation!

I have a great respect for Roger Bivand, but I believe his suggestion of a list server as a means of communication is one of the topics where reasonable people can disagree.

I believe that a lot more happens on Stack Overflow and https://gis.stackexchange.com than on the list server. The number of people willing to answer questions (and by extension the quality of the answers) over there is formidable.

I will leave the issue of possible spatial topic creation to the RStudio folks, as it ultimately is them who pays the bills; but should a spatial topics be created I would gladly subscribe to it & help with the newbie questions over there (which is what I do anyhow, so not much extra commitment on my side :slight_smile: )

5 Likes
#10

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.