deducer and shiny?

[Vincent]

I saw the application linked below built using Shiny (by Rstudio). The concept reminded me of plotbuilder in terms of how interactive it is. 

I used Deducer and JGR in class this quarter and students were very positive (Thanks Ian!!!!). However, when there were problems it was always with JGR. 

Could Shiny be a way to bring Deducer to a wider audience perhaps? For smaller datasets interactive graphing and analysis would be great for students. For larger datasets the interactive ('reactive') feature might slow things down. However, you could (presumably) turn this feature off as well.

http://glimmer.rstudio.com/winston/heightweight/

[Ian Fellows]

Shiny runs on a completely different architecture. I don't see any way
to meaningfully meld Deducer and Shiny, but I'm always open to being
proven wrong...

best,
Ian

[Vincent]

This is only a 'proof-of-concept'. The below may not work from JGR.

install.packages('shiny')

library(shiny)

shiny::runGist('4385800')

Code is on github

https://github.com/mostly-harmless/damshiny.git

[Ian Fellows]

I get

> library(shiny)
> shiny::runGist('4385800')

Error in getOct(block, 100, 8) : invalid octal digit

ian

[Vincent]

Apparently GitHub recently updated their gist service and that broke the runGist function in shiny. The put a new version of shiny on CRAN around December 18th. Can you try updating your version of shiny? It works on 0.2.4.

[Ian Fellows]

Yup, it works with the updated version. I like what you have done
here, but I would caution against going too far down this road. Shiny
was not built as a general purpose toolkit appropriate for a full
scale data analysis GUI. From a classroom perspective though, I can
see it being very useful, allowing students a more restricted
environment to explore predetermined data in a structured way.

best,
Ian

[Vincent]

Thanks for the input Ian. I am designing this for my students.