Is Deducer still being maintained?

[Bruce Miller]

Hi Ian et all,

Are you still maintaining and updating Deducer?

Seems several plot options are no longer working with the most recent
versions of R.

Bruce

[Anthony]

I think the answer is no. Not even possible to download it on Windows from my end any longer. My alternative has been Stata as the dropdowns are similar. 

--



---

You received this message because you are subscribed to the Google Groups "Deducer" group.

To unsubscribe from this group and stop receiving emails from it, send an email to deducer+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/deducer/70d97b21-abc4-757b-de23-2c2bf2d3fb2a%40gmail.com.

[CKC]

It was such a great tool for multivariable interactive brushing/data exploration. Sad that it is dead. 

To view this discussion on the web visit https://groups.google.com/d/msgid/deducer/CAK7ekpCHsR0XNwouO2YcFoOpgh%3DqOAk7LN4oVQi8zPWV2JgZvw%40mail.gmail.com.

[Ian Fellows]

Hi Folks,

Thanks to you all for keeping the interest alive. It has been so gratifying that Deducer has been of use to so many people from around the world. When I started coding on Deducer some 13 years ago now, I couldn’t have imagined what a large project it would be or how many people it would reach. Looking back at the code base, you can actually track my progress from a neophyte programmer as I built and learned.

It was fantastic to see others build packages based on the plug-in system I developed, adding functionality and new analytic capabilities.

I have mixed feelings about Deducer going forward. While I had thought 13 years ago that all of the kinks of working with Java from R would be worked out relatively quickly, quite the reverse happened. Java is if anything more difficult to ensure is set up correctly with R and JGR is also not being actively maintained. For a tool that is geared toward less sophisticated users, this is a critical technological gap. Further, with the rise of analytics on the web and RStudio, it is clear that desktop applications like Deducer are not where the future (or even the present) is.

There are aspects of Deducer that are being actively maintained. In particular RDS Analyst is used around the world to estimate HIV prevalence in hard to reach populations, and the CDC regularly runs trainings in it with in country partners. More broadly, I’ve struggled in recent years to provide the ongoing development necessary to have Deducer run perfectly. In particular, the windows installer and plot builder tend to get out of date rather quickly. Running a company requires a great deal of effort, leaving little time for passion projects. 

That said, with some exceptions around the plot builder, Deducer still runs and works great so long as your R and Java are set-up correctly.

Ideally, Deducer wouldn’t depend on a single developer, but unfortunately we have not attracted anyone to build/maintain the core of Deducer. I’d very much welcome and support anyone who wants to take up the mantle. The code is of course open source and available at https://github.com/ifellows/Deducer.

I had begun work on a successor to Deducer called Vivid ( https://github.com/fellstat/vivid https://www.youtube.com/watch?v=fJMhS2BUf_c&feature=youtu.be ), but progress on that has stalled as we wait for funding and development resources.

So that is where we are right now. I welcome any suggestions that the community has on a path forward.

Best,

Ian

To view this discussion on the web visit https://groups.google.com/d/msgid/deducer/CAHZNfbpCwhYUDNApDn7HyGR39MG51D9Vnn_ufg9WJ%2BbvpK%3DYZg%40mail.gmail.com.

[Rudolf]

Thumbs up. I think Deducer ist still an incredibly useful tool, which I would not want to miss. Just keeping the status quo would be good enough. It may not be perfect and take too much effort on Ian's part to bring it at par with a truly modern tool, but it does still work well and it meets so many basic needs of the user. It is also expandable with own GUIs that are not too hard to make. Hence,, a few fixes to keep it up and running here and there would be all what is required and much appreciated.

For me, it still runs perfectly well on the Mac, as of today, with all the latest upgrades, namely MacOs Big Sur, R 4.0.3, Oracle Java JDK 11 and higher, rJava 0.9-13, JGR 1.9-1.  Pls note JGR 1.9-1 is the Github version, kindly made available by the developer. The CRAN version of JGR is not compatible with more recent Java versions from Oracle.

[Bruce Miller]

Hi all,

Some years back I recall there was a way to display the R code used in
Deducer so one could replicate it when needed to recreate the same
analysis with different data sets.

I can not find where that is now in the latest version.

I have looked under Window and Object Browser but can not find this.
Any suggestions welcomed.

Bruce

[Rudolf]

Isn't the code printed out to the console, so it would just needed to be copied and the variable and data term be replaced by the new ones.  

Example:  descriptive.table(vars = d(newvariable),data=newdata,+ func.names =c("Mean","St. Deviation","Valid N"))

After opening Plot Builder you can also go to Menu and select View call to look up or copy the respective code.

[Ian Fellows]

Exactly right. You can press the up button to cycle through your call. Additionally, you can select save, and save just the commands to a file.

Best,

Ian

--

---
You received this message because you are subscribed to the Google Groups "Deducer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to deducer+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/deducer/45b1bb96-c233-4d6b-a5ed-287d8432135cn%40googlegroups.com.

[Bruce Miller]

Hi all

I have been getting results in Deducer that do not display correctly. 
Seems the font has changed?
See below.
Bruce

$` ÿþSex: Female ÿþ`
      ÿþValid N.Forearm ÿþ       ÿþMinimum.Forearm ÿþ
            64.000000             32.500000
      ÿþMaximum.Forearm ÿþ          ÿþMean.Forearm ÿþ
            40.200000             37.381250
ÿþSt. Deviation.Forearm ÿþ
             1.620589

$` ÿþSex: Male ÿþ`
      ÿþValid N.Forearm ÿþ       ÿþMinimum.Forearm ÿþ
            30.000000             33.300000
      ÿþMaximum.Forearm ÿþ          ÿþMean.Forearm ÿþ
            38.400000             36.230000
ÿþSt. Deviation.Forearm ÿþ
             1.514789

[Bruce Miller]

Deducer not displaying font correctly

Hi all I still rely on Deducer to run most repetitive things.
However aver some of the more recent updates of R and re-installing
Deducer I am having display issues.
The results form some procedure are displaying leading and trailing
characters for labels.
How can I fix/ remedy this?

getLowCV(ptemes)
[1] " ÿþFmin ÿþ"  " ÿþFmax ÿþ"  " ÿþFmean ÿþ" " ÿþFk ÿþ" " ÿþFcH1 ÿþ"
[6] " ÿþFc ÿþ"    " ÿþFcH3 ÿþ"

[Bruce Miller]

Hi all and Ian,

I continue to use Deducer as it is very versatile and helpful vs R Studio for my use.

I have had an issue with eh console display showing symbols rather than text and changing font does not remedy this.

How can this be fixed?
Example below.
Best
Bruce

[Rudolf]

If it helps, this is not an issue with Deducer. 

Deducer still works well, even after upgrading to the latest R release, 4.2.

No problems with this displaying correctly here. I am on macOS 12.4 though.

Perhaps set your Windows system to UTF-8.