Question from community to @Derek_Ouyang_Stanford
*When it came to choosing Shiny for your dashboard platform, was this chosen just for convenience; did you consider other options like Tableau, or Plotly/Dash, or Esri or any other platforms?*
*Would you recommend it to other government agencies / collaborations? Does it have any drawbacks for your end-users?*
First I’ll distinguish the processing we do in R (where our preference is R for a variety of R-specific reasons) from the dashboard design we’re choosing to do in R Shiny, for which we certainly could have made other choices.
• We do have a captured effect of dashboard design being easiest in the language we were doing the data processing in, which certainly makes things easier but doesn’t necessarily mean we have the best possible dashboard.
• In the case of a separate project on mapping food distribution resources, we have made initial mockups in Shiny, but then handed it off to a development team that built out a more sophisticated tool in javascript and with Mapbox as the mapping engine here. Besides aesthetics, the biggest concern was the processing speed of Shiny on older phones. We certainly agree that for something with much larger traffic, Shiny would have performance limitations, but our main use cases have been for smaller audiences.
• We haven’t looked too deeply into Tableau or Dash but would generally characterize Shiny as having more customizability, and we’ve found that useful in being able to react to our partner’s requests. That being said, we are using Plotly charts and generally find them to be the most versatile in the world of interactive R plots.
• With Esri, we have primarily taken issue with the barrier to customizability and bespoke processing, which is doable in Esri but then relies on either javascript knowledge which is not the focus of our initiative, or ArcPy which we generally have not found to be efficient or accessible (i.e. requires license).
For folks doing the kind of work we’re doing, where we have a long-term relationship with local governments, I think R Shiny is a good balance of teachable and versatile. But for in-house government work, I see most of that is still probably best done using more corporatized tools like Esri ArcGIS Hub for open data platforms, where the focus is more on data maintenance and interoperability.
Super helpful insights, thanks for sharing your experience