R can be used for big data analytics by integrating it with Hadoop via packages like rmr that allow R code to be run on Hadoop clusters using the mapreduce programming model. This exposes the mapreduce API within R and hides the complexity of Hadoop. Other R packages provide interfaces to higher level frameworks built on Hadoop like Hive and Pig. The document provides examples of using R and rmr to perform mapreduce operations like kmeans clustering on large datasets stored in Hadoop. It also shows how a kmeans algorithm can be implemented in Pig Latin and integrated with R through a Java UDF.