R/Shiny vs FileMaker (discussion about data visualization and reporting)

[Bobby Rohrkemper]

I work at a company that has about 20 years invested in FileMaker development. We're a marketing company in the tourism industry. We have very little automated reporting and no automated data visualization. I have experience with R and Shiny and would like to start using this experience to do analysis, create charts, and to automate reporting.

I'm looking for some good references that describe the advantages and disadvantages of these platforms:
- FileMaker (relational database with integrated forms, reporting, basic charts)
- R and any related products such as Shiny, R Markdown, RStudio Server, Shiny Server, shinyapps.io, etc

Here are some specific questions, but any information would be welcome, especially general arguments that can be presented to a non-technical board of directors.

- Which types of tasks/analysis should be done in our relational data base (FileMaker) and which are better done with R?
- What can be done with R that would be tedious or impossible with FileMaker?
- How are the graphics in R better than FileMaker (in my option, FileMaker charts are similar to Excel but perhaps less customizable)
- How can we best set up our infrastructure so that users aren’t overwhelmed with knowing when they should go to FileMaker to find what they’re looking for and when to look at a Shiny webpage? Should we imbed the Shiny page within a FileMaker database as a separate layout?

I'll be checking back to answer any follow-up questions.

[Bobby Rohrkemper]

To clarify, our plan is to host the data and to do analysis in R. Some basic analysis can also be done in FileMaker. And some kinds of data management tasks can be done in R. My question is about where we should set this boundary. I'm looking for both theory and practical examples.

[Joe Cheng]

I have not used FileMaker Pro in over twenty years so it's not really fair for me to draw comparisons. In fact I'd be very surprised if the overlap between FileMaker Pro users and R users is very large at all, as (correct me if I'm wrong) FileMaker Pro is primarily aimed at businesspeople at small businesses or small workgroups within larger businesses, while R is aimed squarely at statisticians and data scientists.

• Because R and Shiny are primarily code- and text-file based (rather than a binary file format like FileMaker):
• Scripts and apps are reproducible, can be version controlled, and can easily be reviewed by others. For me, this alone would make me shy away from FileMaker for anything that's important to get right; the idea of building apps without version control is frankly terrifying to me. It's not just about backups (which you can obviously do with FM or anything else that uses files), it's about knowing what changed when, and being able to review what you've changed before you make a commit.
• You can easily set up test and staging environments to review your apps on non-live data before committing changes to production.
• You can have two people working independently on the same app's code, without disturbing each other, yet their changes can easily be merged together (automatically or manually) whenever each is ready to commit those changes. (i.e. git branch)
• You have to know how to code. You have done some R so you know it takes patience (and talent, to some degree) to learn how to program proficiently. This is too high a barrier of entry for tons of people out there, especially if their databases/apps are not important enough to them to make a big time investment.
• R and Shiny are designed for modularity and re-use (at the function, script, and package levels). This means it's very easy to leverage solutions that others in the community have come across, and this particular community has an extremely deep-rooted culture of sharing. (See the many thousands of packages available on CRAN, BioConductor, R-Forge, and GitHub, compared to the few dozen FileMaker Pro modules on modulefilemaker.org.)
• Innovation in R is primarily driven by the community, not by the small group of people in R-Core. I imagine that as a FileMaker user, if there are new features you want to see in FileMaker, you generally have to wait for the company's engineers to implement them.
• I don't know enough about FileMaker to provide a detailed breakdown of the kinds of charts you can make in R vs. FM, but I would be stunned if FM was even close to as flexible as R's base graphics, ggplot2, and/or lattice. A google image search of "R graphics" should give you an eyeful. In particular, ggplot2 strikes a really appealing balance, being high-level enough to be very pleasant to use but low-level enough that the universe of chart types you can make with it is enormous. And there are newer packages like rbokeh, plotly, and ggvis that let you work with similar grammars but give you interactive visualizations instead of static ones.
• I really don't know anything about statistical modeling, forecasting, diagnostics, etc. but I'm told R is really good at them. :)
• R can work with many common databases and file formats but I'd check to make sure it can work with FileMaker. If not, I imagine you'll be doing a lot of importing and exporting using CSVs and that doesn't sound like fun.

There's a lot LOT more than R can do that I haven't gone into, but hopefully that provides a start.

--
You received this message because you are subscribed to the Google Groups "Shiny - Web Framework for R" group.
To unsubscribe from this group and stop receiving emails from it, send an email to shiny-discus...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/shiny-discuss/2f11f26a-90c2-4847-afc6-248e93092c4c%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Bobby Rohrkemper]

Hi Joe,

This was a great reply which helped to bring some perspective. I've also been speaking with some of your colleagues at RStudio and they pointed out some of your Shiny demo apps. Thanks very much.

Cheers,

Bobby

On Tuesday, February 23, 2016 at 9:23:13 AM UTC+1, Bobby Rohrkemper wrote: