I am trying to collect a few examples of shiny dashboards that are hosted on a .com or .gov domain (i.e. Shiny being used in production). I am trying to convince the senior management of my company to seriously consider Shiny in production for analysis and insights delivered to clients. We currently use Tableau and an internally developed tool. Ideally, I am looking for examples in Finance, Retail, Manufacturing, Government, etc.
Over the past few months I have collected these examples:
Thanks @slopp ! This is a fantastic resource. I had seen it before, but it looks like there are more case studies there now. I will definitely look through this in more detail.
From the UK but, if it helps, we use shiny (and plumber) in production (finance industry). The app linked below is mainly marketing focused but we do also have other apps which are limited to specific clients. https://tools.hymans.co.uk/tprfasttrack/
If you can, let us know how it goes.
Tableau and Shiny are such different tools so I'm really interested to hear their feedback and their perceptions of the tools.
Thank you @jcolomb ! The WHO Shiny Dashboard is quite impressive indeed.
@lxy009 I am happy to post updated on this. At the moment, two of sticking points are data security and cost. Tableau makes "push button publishing" quite easy for enterprise applications. I guess R Studio Connect is the closest equivalent to this. But our IT team does not have much knowledge about setting up Shiny servers. So, at the moment, we are trying to look up some examples of other public Shiny dashboards.
That is running on shiny server but for internal apps we use shinyproxy. It's open source, uses docker for deployment and has lots of enterprise authentication options, so maybe your IT team would be more familiar with these technologies.
At BC Stats, the Province of British Columbia's statistics agency, we are using Shiny for both data visualization and data dissemination. Here's three examples--we have some others that are not yet published.
Note that in all cases they are running on RStudio's cloud service shinyapps.io; the code is up on the BC Government GitHub site, and the data we are using for these is under various open data licenses. I have been working on getting Shiny server or RStudio Connect, which would provide a significant expansion in functionality, but I keep running up against various institutional barriers.
Data visualization
BC Housing Market - combines Census and administrative data, with a variety of plot types and maps
BC Student Outcomes - displaying results from surveys of post-secondary students after graduation
Data dissemmination
Population estimates - this Shiny app provides access to population estimates for a variety of geographies, age categories, etc. The Shiny app is a straight copy of an older tool we had, designed for self-serve data downloads. The next phases of development will be to incorporate some feedback we have heard from clients, including creating some data visualizations similar to what's in the housing app. (Note that there are parallel apps for household estimates and population projections.)
I'd be happy to chat further if you want more information and/or a pep talk the next time you run up against institutional barriers!
Interesting that data security and cost are the sticking points since, in my opinion, those are not Tableau's strong suits. For what it's worth, I've set up both Shiny servers and Tableau in parallel at the same company. Tableau more for self-service analytics case and where we spent the time to curate the data and expose the right elements ... and where the analysis was simple and straight-forward. Shiny was used when data still required a lot of manipulation not easily or impossible to be done on Tableau and where there's modeling and analysis; things that are not appropriate for self-service.
The 3rd factor, IT not having experience with a software so not wanting to do it, is a common one. I've always been willing to get my hands dirty and be responsible for administrating and working with IT, so I haven't had as much push back in that regard.
Hi Richard, thanks for your input. Would you be able to elaborate on why security is not one of Tableau's strong points?
As for the 3rd factor - you are exactly right. The team has has specifically called out security and costs as primary hurdles. However, I suspect that lack of familiarity is probably an even bigger hurdle. They just haven't haven't verbalized that barrier clearly.
With that point in mind, I had posted another question on this community (here: Talking to "traditional" IT managers about Shiny). If you have any other resources that I can utilize, please let me know!