Role of R Studio Cloud


#1

I just saw R Studio Cloud and am curious about how folks see it fitting into the R / R Studio suite of tools. I can see it being useful for getting people up and started for teaching and learning R and, maybe (and depending on how the accounts work / are limited), running analyses that take awhile to run.


#2

Maybe consider moving this thread to the #rstudio-cloud category?


#3

I think there are a lot of different ways in which we envisioned rstudio.cloud fitting into the suite of tools. There are two core use cases we wanted to tackle in the first iteration:

  • Simplify the life of a learner by making it easy for them to hop on and play with R.
  • Simplify the life of a practitioner by making it easier for them to share working examples with others.

There are many other use cases we could go after, but we would frankly like for those use cases to be informed by real demand/usage.


#4

hi @tareef, I see the cases that you describe. On the other hand I would love to have a service comparable to a 16 GB RAM local PC but in the cloud. Say, to do large calculations but not huge. Too demanding for my old noteook but doable for high end notebooks and medium desktop. Do you think you will be able to provide such a service (maybe in the paid version)? This would be my case. Shifting my hardware cost to service cost ... but I guess this is obvious :slight_smile:


#5

Hi @RichiW, That is definitely a use case that we want to support. We are still in Alpha and are trying to learn about the use cases for which people would want to use rstudio.cloud. I happen to agree with you that it would be powerful to have that option. One question we have often struggled with is whether folks would want a "pay as you go" model, or whether they would prefer a monthly fee that gives them a fixed amount of power/compute.

Thoughts?


#6

@tareef: I think in my case I would prefer the "pay as you go" version. There will be times when I do hyperparameter tuning and feature engineering - then I need massive power. When it is all settled (at least for a while) then I would need much less power doing training and prediction.

Thus logging on and finding my session (at least all the code and data) as it was before with big ressources one time and with less ressources another time would fit my needs best.

Another question: I can open rstudio.cloud on my android phone but I don't see how I can start just a single command. Should this be possible? In the far future I could imagine an android app :slight_smile:
Thank you for all the work here. This could be the service I was looking for for so long!


#7

I totally agree with RichIW- I recently left Digital Ocean because of their monthly fee policy. My work in R is periodic and not regular and it becomes VERY irritating to get demands for payment each month when you haven't used the service. Pay as you go is the way to go!
I also support the second comment of RichIW- it would be wonderful to have access to serious computing power, as with Azure or AWS directly with/through Rstudio cloud. If this was feasible easily then Rstudio cloud would become my only "computer".


#8

My apologies @RichiW and @pushka for the delayed response. I hear you loud and clear on the pay as you go. Our initial thinking with shinyapps.io was that we would make it a pure pay as you go model, and our first 20 conversations with prospects and friendlies all came back with the same general response: "I personally would prefer a pay as you go model, but my Finance department needs predictability and as such we would want/need a more committed model that is more predictable". That being said, I think rstudio.cloud is a bit of a different animal precisely because of the need to burst for compute. I suspect we will need to support both models on rstudio.cloud. Thank you both for your feedback. It is incredibly helpful to get your perspectives.


#9

Hi,
I wondered if there is any intention to enable companies to host the rstudio cloud environment on their own preferred infrastructure. The potential for sharing and collaborating looks attractive (albeit I confess to not having spent too long yet) but from an enterprise perspective there will be data, security and privacy issues. So having an RStudio Cloud equivalent to RStudio Connect could be quite appealing?
Thanks,
Nick


#10

I use r cloud for 2 reasons

  1. To share my code to other people who can help me out provide solutions in sites like these.

  2. For beginners who don't have access to r studio due to IT policy in my office. I suggest them to start with r cloud and learn it.


#11

For that RStudio Server Pro is a good option.