Today’s idea is use Quandl approach in other applications. Jenkins exports some of its objects to its external JSON API. And there are stats being collected from plug-ins (like the download statistics). I am curious to know if there would have any gain in exporting parts of this data as R matrices and processing it directly with R.
This could be used in a lot of different applications, perhaps even using Hadoop and R. Instead of writing collectors to consume an RESTful service, you would simply write one R line to get the job done.
R does very fine in plotting too, so I wouldn’t be surprised to at least see some prettier charts :o)