3

I would love to get some opinions about those three tools.

I am working with R and Shiny quite long time. I developed many apps (production side) which are actually being very popular at my company. The new task came from the "up" to create the KPI/BI dashboard and here those three names came up: Shiny R / QlikView /SAP Lumira. After researching on the web I am still standing behind R Shiny (I do not have any experience with QlickView or SAP Lumira), however it seems that another colleagues from the marketing etc. part are not very much convinced.

Any opinions about:

1.Live data visualisation in QlikView and SAP Lumira

2.calculation of specific parameters such as Cp and Cpk values

3.Advanced plotting dependend on the data (for example line plot with specific interactive horizontal lines which are calculated tolerances)

4.Reports

5.Connection to SAP from R Shiny/QlikView

6.Creation of dashboards

would be highly appreciated

Ferdi
  • 4,882
  • 7
  • 42
  • 62
Mal_a
  • 101
  • 7
  • Have you read the official statement of RStudio on the topic: https://support.rstudio.com/hc/en-us/articles/218294727-Why-would-I-use-Shiny-instead-of-Tableau-Spotfire-Qlikview-or-similar-BI-tools- – Ferdi Feb 24 '17 at 08:40
  • 1
    This might appear of topic on the first glance, but I suggest to leave it open because of the high relevance to statistical practice. – Ferdi Feb 24 '17 at 08:41
  • 1
    I actually read it before, and those arguments are not enough to convince the "marketing department" which does not have much R experience or so. I would like to get some opinions from the people that had a chance to work with some of these tools – Mal_a Feb 24 '17 at 08:42
  • I have the same problem. ;-) – Ferdi Feb 24 '17 at 08:44
  • 1
    Good to know that i am not alone in that! From this what i have researched and read, Shiny is the best solution:price (we do own at the moment Shiny Pro), flexibility (however it is hard to bring up strong arguments to the table when next to you are giants such as QlickView and SAP giving support, etc). It is actually hard to explain to them that Shiny is not actually a "ready product" as QlikView and SAP... – Mal_a Feb 24 '17 at 08:50

2 Answers2

4

Disclaimer: I am an extensive user of R Shiny and R in general and I have not used any of the other above mentioned tools (Tableau, QlikView, SAP Lumira). In this post I would like to outline some of the major benefits/differences of Shiny compared to the other tools.

  1. PRICING

RShiny comes for free. However some features cost extra. The RStudio Shiny Server Pro Licence costs \$9995 per Year for 20 users. RShinyPricing Deploying the apps may cost extra if you are looking for special feature like premium support or custom domains. Tableau Tableau costs you at least \$999, but it comes for free if you are a student or an instructor. Tableau pricing. A free trial for limited time is available Qlikview Qlikview has a free version, but the free version is very basic and you have to buy an upgrade to use it professionall QlikView pricing SAP Lumira Similarly to Tableau you can have a free trial. However if you use it permanently you have to pay 185$ per User. SAP Lumira Pricing

RStudio clearly has the better pricing. Freeware is always cheaper than expensive commercial products.

  1. DataImport

The BI Tools have access to many different data sources. Among other tableau can import data from SAP, Excel and Salesforce. Also QlikView can import data from many different sources. Most commonly people import data from Excel, CSV or SQL Server. Also SAP Lumira can import data from many different types of data, e.g. excel, CSV and SAP. However RShiny can take its data “directly from R”. You can use all the data import libraries on CRAN or Github and you can use your own onces. The dataimport libraries in R include RODBC (for relational databases), foreign (e.g. for Stata files and SAS files), xlsx (for excel files), Neo4j (for data from graph databases) and readr (for many different datatypes, e.g. tsv files). Here is a library which imports data from SAP directly to R(Shiny). I have never heard of any of the other tools importing data from xlsx, Stata, SAS or Neo4j. Therefore RShiny is the clear winner In dataImport.

  1. Calculation of specific statistical values

Well, R Shiny is part of R and R is the statistics language with the most different algorithms implemented. This points clearly goes to RShiny.

  1. Advanced data visualisation

RShiny can be combined with all other data visualization libraries: ggplot2, htmltable, igraph, plotrix and lattice just to name a few. Note that many JavaScript libraries (e.g. plotly.js, d3, Datatables and dygraphs) are also implemented in R. Have a look at the htmlwidget library. You can also use extensions of RShiny such as shinyjs or you can include html, css and JavaScript easily.

Therefore RShiny offers more flexibility concerning data visualization.

  1. Steeper learning curve

RShiny might have a steeper learning curve, but if you are good at RShiny you also learn programming and it will be easier to learn other programming languages especially functional programming languages such as F# or Closure.

  1. Reproducability

Anyone can take your app with the code install RStudio and deploy it. As far as I know drag- and -drop tools such as Tableau do only offers very sparse options concerning Reproducability.

Ferdi
  • 4,882
  • 7
  • 42
  • 62
3

I use reporting/BI tools. Didn't use QlikView especially, but I think the main concept and advatage of QlikView (and Tableu, ...) is that you load a lot of data into the tool, create a dashboard/report and then you can do live-filtering (selecting something in 1 table updates all other elements). Also you can distribute these reports in your company, by email or by specifying access-rights (which data / report can be accessed by which user group).

On the other hand R allows you to build custom coded visualizations / reports which answer 'exactly' the business questions. Most BI tools allow some kind of scripting, so you can compute KPIs there as well.

So I think this is a trade-off between R being more flexible and BI tools more comfortable and easy to integrate in the enterprise. Somehow everything is usually possible in BI tools, but if you don't use the standard visualizations, then you soon get to 'creative workarounds'.

Probably you should do a list with the most important feature demands from your business experts and check those in the different tools (or discuss them here).

As a compromise you might recommend a BI tool which allows R integration like Microsoft Power BI.

Tobi
  • 111
  • 4