googleVis 0.3.2 is released: Better integration with knitr

2 comments
After last week's kerfuffle I hope the roll out of googleVis version 0.3.2 will be smooth. To test the water I release this version into the wild here and if it doesn't get shot down in the next days, then I shall try to upload it to CRAN. I am mindful of the CRAN policy, so please get in touch or add comments below if you find any show stoppers.

So what's new in googleVis 0.3.2?

The default behaviour of the functions print.gvis and plot.gvis can be set via options().

Now this doesn't sound too exciting but it can be tremendously helpful when you write Markdown files for knitr. Here is why:

2 comments :

Post a Comment

googleVis 0.3.0/0.3.1 is released: It's faster!

2 comments
Version 0.3.0 of the googleVis package for R has been released on CRAN on 20 October 2012. With this version we have been able to speed up the code considerably. The transformation of R data frames into JSON works significantly faster. The execution of the gvisMotionChart function in the World Bank demo is over 35 times faster. Thanks to ideas by Wei Luo and in particular to Sebastian Kranz for providing the code.

Further, the plot function plot.gvis has gained a new argument 'browser'. This argument is passed on to the function browseURL. The 'browser' argument is by default set to the output of getOption("browser") in an interactive session, otherwise to 'false'. This prevents R CMD CHECK trying to open browser windows during the package checking process. Unfortunately this caused an error message under RStudio and RGui on Windows. The argument has been removed again and plot.gvis handles the check if R is running interactively internally. The bug has been fixed in googleVis-0.3.1, not yet available on CRAN, but on our project download page. Thanks to Henrik Bengtsson and Sebastian Kranz for their comments, suggestions and quick response.

2 comments :

Post a Comment

From guts to data driven decision making

4 comments
Source: Wikipedia, License: CC0
There is a wonderful cartoon by Loriot, a German humorist (1923 - 2011), about a couple sitting at a breakfast table, arguing about how to boil a four-and-a-half minute egg. The answer appears simple, but husband and wife argue about how to measure the time using experience, feelings and expert judgment (wife) or a clock (husband).

The whole sketch is hilarious and is often regarded as a fine observation of miss-communication.

Yet, I think it really points out two different approaches in decision making: You can trust your guts or use data/measurements to support your decision.

4 comments :

Post a Comment

Review: Kölner R Meeting 5 October 2012

2 comments
The third Cologne R user meeting took place last Friday, 5 October 2012, at the Institute of Sociology.

The evening was sponsored by Revolution Analytics, who provided funding which went towards the Kölner R user group Meetup page. We had a good turn-out with 18 participants showing up and three talks by Dominik Liebl, Jonas Stein and Sarah Westrop.



Photos: Günter Faes

2 comments :

Post a Comment

Connecting the real world to R with an Arduino

3 comments
If connecting data to the real world is the next sexy job, then how do I do this? And how do I connect the real world to R?

It can be done as Matt Shottwell showed with his home made ECG and a patched version of R at useR! 2011. However, there are other options as well and here I will use an Arduino. The Arduino is an open-source electronics prototyping platform. It has been around for a number of years and is very popular with hardware hackers. So, I had to have a go at the Arduino as well.

My Arduino starter kit from oomlout

The example I will present here is silly - it doesn't do anything meaningful and yet I believe it shows the core building blocks for future projects: Read an analog signal into the computer via the Arduino, transform it with R through Rserve and display it graphically in real time. The video below demonstrates the final result. As I turn the potentiometer random points are displayed on the screen, with the standard deviation set by the analog output (A0) of the Arduino and fed into the rnorm function in R, while at the same time the LED brightness changes.

3 comments :

Post a Comment