Support website integration

[clnsmth]

What would it take to integrate our app into a website? This feature was originally set as a post-production-release enhancement.

[wetlandscapes]

Here are some common methods for running shiny apps on the web:

• Via RStudio: https://shiny.rstudio.com/deploy/
• Via Digital Ocean: https://deanattali.com/2015/05/09/setup-rstudio-shiny-server-digital-ocean/
• Via AWS: https://medium.com/@CharlesBordet/how-to-deploy-a-shiny-app-on-aws-part-1-4893d0a7432f
• EDI/LTER/DataOne server: Would need to talk to IT and the system administrators running those sites.

The RStudio and DigitalOcean options are appealing, because they are easy to implement (more so the RStudio option). However, that could potentially get expensive as more instances and time on the app generally mean more money. There are academic discounts, however: https://www.rstudio.com/pricing/academic-pricing/. That said, one could spend \$3,000-20,000/year on a shiny stuff. For specific info on hosting from shinyapps.io (\$3,300/year): https://www.shinyapps.io/

A seemingly cheaper option, but one that would require more back-end support from IT, is to use a server-hosted container system (e.g., Kubernetes and Docker) to dynamically spin up instances of R on a server that could host the Shiny app, dependent on demand.

Once the app has initialized its simply spewing out html, css, javascript, etc. That means it could be "placed" into an existing website using an iframe. That is, the web design crew responsible for the website would have to explicitly write in code to include instances of the shiny application via some host server (listed above).

Conclusion:

• Simplest, but most expensive option is to use RStudio commercial products.
• Cheapest, but most complicated is using a self-hosted server with containers, which would have to be run by an IT department somewhere.

[CoastalPlainSoils]

With An's question in another thread about processing large datasets, would both of these options run at the same speed or would one be faster than the other? @wetlandscapes