In this post, a way to overcome one of R's limitations of lack of multi-threading is discussed by job queuing using the jobqueue package

Run SparkR in Hive Context to reach the Hive UDFs and window functions the SQL Context lacks, compared against dplyr, plus building Spark with Hive.

Run SparkR in local and cluster mode from an R project, setting environment variables and the library path, then reading JSON and CSV files.

Set up a two node Spark standalone cluster on VirtualBox Ubuntu guests, covering machine preparation, copying the VDI image and password-less SSH.

Wrap Python boto calls for Amazon S3 in an R package, rs3helper, where the Python scripts return JSON so R parses it as vectors and data frames.

Call a SOAP web service from Python with the suds library, a job R has no comprehensive client package for, demonstrated on the Sizmek MDX API.

Write Python in an object-oriented style, shown with SOAP API client classes, and a reading order for the Introducing Python book by chapter group.

Setting up random seed is important for reproducibility of analysis. In this post, we discuss how to generate random seed using the caret package.

Turn an analysis into an R package, with the treebgg bagging package as the worked example, covering roxygen2 documents, testthat tests and vignettes.

Compare the snow and foreach approaches to parallel processing in R on three practical examples, starting with k-means clustering on the Boston data.