Dynamic visualisation of R data using Adobe FLEX

Sat, May 2, 2009 12:07 AM #

Hi Harsh,

If your project is so specific on the regression model, my suggestion
is to use tcl/tk or Gtk2 interface instead of FLEX, because your
requirement on interaction is simple (e.g. you don't need to
drag-and-drop points).

I think the example 'run.cor2.examp()' in 'TeachingDemos' package can
give you enough inspiration.

Or if you want to do the job (to change the slope) automatically, you
may want to see the example in the 'animation' package:

##
install.packages("animation")
library(animation)

# default animation in R: with slope changing
least.squares()

# animation in an HTML page
oopt = ani.options(ani.height = 450, ani.width = 600,
    outdir = tempdir(), nmax = 50, title = "Demonstration of Least Squares",
    description = "We want to find an estimate for the slope
         in 50 candidate slopes, so we just compute the RSS one by one. ")
ani.start()
par(mar = c(4, 4, 0.5, 0.1), mgp = c(2, 0.5, 0), tcl = -0.3)
least.squares()
ani.stop()
ani.options(oopt)
##

P. S. I'm quite interested in interactive visualization using R and
Flash (you may also want to know Flare and SWF Tools!), so please let
me know if there's any progress in your project.

Regards,
Yihui
--
Yihui Xie <xieyihui at gmail.com>
Phone: +86-(0)10-82509086 Fax: +86-(0)10-82509086
Mobile: +86-15810805877
Homepage: http://www.yihui.name
School of Statistics, Room 1037, Mingde Main Building,
Renmin University of China, Beijing, 100872, China

On Thu, Apr 30, 2009 at 2:54 AM, Harsh <singhalblr at gmail.com> wrote:

Hi useRs,

I had posted about Adobe FLEX talking to R for rich visualisation.
Reply from Jeffery Horner contained links to the
revolution-computing.com webpage which had information pertaining to
the Bay Users R group Meetup on Web Dashboards with R.

I have a very specific project that I need to implement.
I wish to use the graphics capabilities provided by Adobe FLEX to
visualise outputs from R.
For example:
I would like to fit a regression model to a dataset in R and provide a
FLEX interface wherein, a user may manipulate a slider which would
increase or decrease
a parameter estimate (between the confidence intervals of the
estimate) of a variable to see the effect on the predicted values.

I necessarily have to use FLEX for the interface.
In trying to make sense of how to go about this task, I have
speculated on the following technologies:
1) Using RSOAP which requires a Python client. I am not sure how I can
implement this in Flex.
2) Using RSOAP provided by the Biocep project. This is a java
implementation and maybe this would suit my requirements.
3) The StatDataML package in R provides the creation of xml files of R
objects. Maybe I can use these XML files as encapsulators of R object
data and pass it on to FLEX.
4) Since R has interfaces that connect to most SQL database servers, I
can convert results of lm objects (summary(lmobject)) to data frames
and push them into the SQL database. FLEX can
then read the table and extract the relevant fields for output.

I am aware that I not a computer programmer which is why I may not be
making a whole lot of sense in describing the above scenarios, but I
am trying to gain as much information and programming skills to
'decipher'
the secret behind using FLEX and R.

I have looked at Jeffery Horner's ?BBPLOT project
(http://data.vanderbilt.edu/rapache/bbplot) and honestly I'm amazed.
Replicating this with a FLEX interface is the closest I can get in
describing my requirements.

Any information/technical sources/tutorials etc in this regard will be
much appreciated.

Thank you for your time.

Regards
Harsh Singhal
Decision Systems
Mu Sigma Inc.
Chicago, IL

2 days later

Harsh

Mon, May 4, 2009 6:58 AM #

Hi Yihui,
In trying to make FLEX and R talk, I have decided to approach it by
using the Rserver JRclient as the middle tier in a three tier
approach.
Tier 1: FLEX Visualisation
Tier 2: JRclient (http://rosuda.org/Rserve/example.shtml)
Tier 3: Rserver

I have yet to begin working on this front, but would like to know what
you think about this approach. I understand FLEX and JAVA communicate
seamlessly and a JAVA client to communicate with Rserver seems to seal
the deal for making FLEX and R talk to each other.

Thanks
Harsh Singhal

On Sat, May 2, 2009 at 12:37 PM, Yihui Xie <xieyihui at gmail.com> wrote:

Hi Harsh,

If your project is so specific on the regression model, my suggestion
is to use tcl/tk or Gtk2 interface instead of FLEX, because your
requirement on interaction is simple (e.g. you don't need to
drag-and-drop points).

I think the example 'run.cor2.examp()' in 'TeachingDemos' package can
give you enough inspiration.

Or if you want to do the job (to change the slope) automatically, you
may want to see the example in the 'animation' package:

##
install.packages("animation")
library(animation)

# default animation in R: with slope changing
least.squares()

# animation in an HTML page
oopt = ani.options(ani.height = 450, ani.width = 600,
? ?outdir = tempdir(), nmax = 50, title = "Demonstration of Least Squares",
? ?description = "We want to find an estimate for the slope
? ? ? ? in 50 candidate slopes, so we just compute the RSS one by one. ")
ani.start()
par(mar = c(4, 4, 0.5, 0.1), mgp = c(2, 0.5, 0), tcl = -0.3)
least.squares()
ani.stop()
ani.options(oopt)
##

P. S. I'm quite interested in interactive visualization using R and
Flash (you may also want to know Flare and SWF Tools!), so please let
me know if there's any progress in your project.

Regards,
Yihui
--
Yihui Xie <xieyihui at gmail.com>
Phone: +86-(0)10-82509086 Fax: +86-(0)10-82509086
Mobile: +86-15810805877
Homepage: http://www.yihui.name
School of Statistics, Room 1037, Mingde Main Building,
Renmin University of China, Beijing, 100872, China



On Thu, Apr 30, 2009 at 2:54 AM, Harsh <singhalblr at gmail.com> wrote:

Hi useRs,

I had posted about Adobe FLEX talking to R for rich visualisation.
Reply from Jeffery Horner contained links to the
revolution-computing.com webpage which had information pertaining to
the Bay Users R group Meetup on Web Dashboards with R.

I have a very specific project that I need to implement.
I wish to use the graphics capabilities provided by Adobe FLEX to
visualise outputs from R.
For example:
I would like to fit a regression model to a dataset in R and provide a
FLEX interface wherein, a user may manipulate a slider which would
increase or decrease
a parameter estimate (between the confidence intervals of the
estimate) of a variable to see the effect on the predicted values.

I necessarily have to use FLEX for the interface.
In trying to make sense of how to go about this task, I have
speculated on the following technologies:
1) Using RSOAP which requires a Python client. I am not sure how I can
implement this in Flex.
2) Using RSOAP provided by the Biocep project. This is a java
implementation and maybe this would suit my requirements.
3) The StatDataML package in R provides the creation of xml files of R
objects. Maybe I can use these XML files as encapsulators of R object
data and pass it on to FLEX.
4) Since R has interfaces that connect to most SQL database servers, I
can convert results of lm objects (summary(lmobject)) to data frames
and push them into the SQL database. FLEX can
then read the table and extract the relevant fields for output.

I am aware that I not a computer programmer which is why I may not be
making a whole lot of sense in describing the above scenarios, but I
am trying to gain as much information and programming skills to
'decipher'
the secret behind using FLEX and R.

I have looked at Jeffery Horner's ?BBPLOT project
(http://data.vanderbilt.edu/rapache/bbplot) and honestly I'm amazed.
Replicating this with a FLEX interface is the closest I can get in
describing my requirements.

Any information/technical sources/tutorials etc in this regard will be
much appreciated.

Thank you for your time.

Regards
Harsh Singhal
Decision Systems
Mu Sigma Inc.
Chicago, IL