mages' blog

Next Kölner R User Meeting: 5 October 2012

The next Cologne R user group meeting is scheduled for 5 October 2012. All details and the agenda are available on the KölnRUG Meetup site. Please sign up if you would like to come along. Notes from the last Cologne R user group meeting are available here.

Thanks also to Revolution Analytics, who are sponsoring the Cologne R user group as part of their vector programme.

Using R in Insurance, Presentation at GIRO 2012

Every year the UK’s general insurance actuarial community organises a big conference, which they call GIRO, short for General Insurance Research Organising committee.

This year's conference is in Brussels from 18 - 21 September 2012. Despite the fact that Brussels is actually in Belgium the UK actuaries will travel all the way to enjoy good beer and great talks.

On Wednesday morning I will run a session on Using R in insurance. It would be great to see some of you there.


I prepared the slides with R, RStudio, knitr, pandoc and slidy again. My title page shows a word cloud about the GIRO conference. It uses the wordcloud package and was inspired by Ian Fellows' post on FellStats.

The last slide shows the output of sessionInfo(). I am sure it will become helpful one day, when I have to remind myself how I actually created the slides and which packages and versions I used.

Connecting data to the real world - The next sexy job?

At last week's Royal Statistical Society (RSS) conference Hal Varian, Chief Economist at Google, gave a panel talk about 'Statistics at Google'. Could he get a better audience than the RSS?

Hal talked about his career in academia and at Google. He reminded us of the days when Google was still a small start up with no real idea about how they could actually generate revenue. At that time Eric Schmidt asked him to 'take a look' at advertising because 'it might make us a little money'. Thus, Hal got involved in Google's ad auctions.

Hal Varian at the Royal Statistical Society conference 2012

Another projects Hal talked about was predicting the present. Predicting the present, or 'nowcasting', is about finding correlations between events. The idea is to forecast economic behaviour, which in return can help to answer when to run certain ads. He gave the example of comparing the search requests for 'vodka' (peaking Saturdays) with 'hangover' (peaking Sundays) using Google Insight.

Interactive web graphs with R - Overview and googleVis tutorial

Today I feel very lucky, as I have been invited to the Royal Statistical Society conference to give a tutorial on interactive web graphs with R and googleVis.

I prepared my slides with RStudio, knitr, pandoc and slidy, similar to my Cambridge R talk. You can access the RSS slides online here and you find the original R-Markdown file on github. You will notice some HTML code in the file, which I had to use to overcome my knowledge gaps of Markdown or its limitations. However, my output format will always be HTML, so that should be ok. To convert the Rmd-file into a HTML slidy presentation execute the following statements on the command line:

Rscript -e "library(knitr); knit('googleVis_at_RSS_2012.Rmd')"
pandoc -s -S -i -t slidy --mathjax googleVis_at_RSS_2012.md 
  -o googleVis_at_RSS_2012.html