What do you use RStudio Connect for?

rstudioconnect

#1

I’m curious as to what the community uses RStudio Connect for. I’m looking at potentially pitching it to my employers as a possibility for a data integration and reporting service, because of how easy the reporting is.

So the title says it all, what do you use RStudio Connect for?


#2

This is a great question! I expect our current Connect customers will weigh in, but in the meantime we have quite a few resources on rstudio.com, and I’d recommend:

www.rstudio.com/products/connect
https://www.rstudio.com/resources/customer-spotlight/


#3

Quick question. We are thinking about the same thing at my office (procuring a tool, potentially Connect) and we want to learn about best practices for running this and other types of tools as part of a portfolio of enterprise data tools. What are the right server specifications for a team of say 5-10 analysts as far as number of cores and what not? Thanks!


#4

Hi,

I really like Rstudio connect and use it to rapidly iterate on shiny apps that I am developing.

The push button deployment really cuts down on my development time, and even enables me to change the app during the course of a live demo. Before using Connect, I would use packrat to manually bundle my project when I was ready to deploy it to our Shiny Server so that I could share it with co-workers. This would actually be quite a time consuming process downloading sources etc and would sometime require talking to the sysadmin to resolve dependencies because the development was done on a mac but deployed on a linux server.

Other benefits of using Connect for me are easily restrict access to certain users, specify a custom url for the application, and roll back to previous versions. The biggest additional benefit is that it dramatically lowers the barrier for new shiny developers to publish an application.

Cheers,

Iain


#5

This is a great question for our sales / solutions engineering team, as the answer depends on your user case. We can direct you towards our evals as well.


#6

In my company we use RStudio Connect to offer some self-services to our R users.
We use RStudio connect server to easily deploy shiny application and share Rmarkdown production, all that at the hand of the R developpers from our RStudio Pro servers. One push via the deploy button and it is live.
Next steps is to use RStudio Connect to deploy R functions as a service via API as new version of RStudio Connect allows it now.

Before that, Rmarkdown productions were not so much used and when it was, reports were sent by mail most often. Now it is published in a suited environment allowing much more possibilities for building report (like database connection).
Shiny apps were deployed locally with port opening and it did not managed user sessions, scaled and ensured proper practices for application deployment. Now shiny apps are deployed easily by the developers in a suited environment that allows user controlled access, R sessions management, scaling and versioning. A huge gain for the R community in our company and great incentives toward better R in production.

If RStudio connect has really made the process to deploy and share easier for all our R users, it is also a win for our IT teams because RStudio Connect deals with a lot of R specific stuffs (package management, automated deployement, multiple R versions, …) and offers a lot of professionnal features (LDAP authentication for user management, Logs, Admin interface, configuration scripts,…).

I think a product like RStudio Connect is to consider on both sides, obviously the R users but also the IT teams. In our case, it really helps to fill the gap between both, between R users’ productions and deployed R applications. Also, we began to use RStudio Connect to help us put Shiny application into production - or at least a more production-like state to be precise.

One last comment: RStudio Connect has an impact on how our R developers are working as it tends to require good development practices (Git, dependencies management, better-designed app, …). I see it as a good opportunity to improve quality of our R production - or at least gives incentives toward that. Some could see that as a drawback because it as a short time cost on development (but obviously a mid-term gain in quality).
A poor quality development app won’t work as smooth as a good quality one. As R developers are responsible for deployment in the RStudio Connect environment, they are responsible for the application, which did not pass through IT. Don’t get me wrong, IT teams helps - and it is my job now, but responsibility is shared and it helps proves that a clean workflow helps a lot when maintaining a deployed service (report or apps). By the way, thanks for this help RStudio team ! :wink: