R Packages

Programming with Big Data in R (pbdR)

I am currently the co-lead developer of the pbdR ecosystem of packages. These are a collection of utilities and services for leveraging R in HPC environments. For more information, see the main project website http://r-pbd.org/

System Hardware Info for R and Analytics (SHINRA)

SHINRA is a set of tools for benchmarkers working with R. For more information about SHINRA, see https://shinra-dev.github.io/

Analysis Packages

Package Description Download
coop Fast, efficient implementations of the “co-operations” (covariance, correlation, cosine similarity). ghbutton
cranbutton
pcapack An experimental package for computing principal components and related statistical quantities in a high performance manner. ghbutton
linmod An experimental rewrite of R’s lm.fit() and glm.fit() functions. Like R’s implementation, it uses the “limited pivoting” strategy in a rank-revealing QR; unlike R, we use level 3 BLAS (and generally better memory management) when we can. ghbutton
yalda “Yet Another LDA” package. Wraps the serial portion of the Google plda program for use with R. ghbutton

Utilities

Package Description Download
dequer A package offering a custom data structure for R. A deque is like a list, but with very cheap insertions (since R lists are contiguous arrays of pointers). ghbutton
cranbutton
lineSampler Samples a file by line; quickly read a subsample of observations from a large flat text file (e.g., a csv). Useful for downsampling; not suitable for resampling. ghbutton
RNACI A collection of utilities that make dealing with R’s native C interface more manageable. Notably the package handles simplified SEXP allocation (with automatic gc counting), and the creation and management of lists and dataframes. For instance, by making use of C99’s va_args, we can create arbitrary lists of named objects in just 2 simple function calls. ghbutton
Rth Multicore programming in R through the thrust template library. Compiles to an OpenMP, Intel Thread Buildingblocks, or CUDA backend. ghbutton
rexpokit A set of bindings and utility functions for computing matrix exponentials via the Fortran 77 package EXPOKIT. ghbutton

Just for Fun

Package Description Download
Rdym A “did you mean?” tool for R to help catch spelling mistakes in an interactive R session. ghbutton
Rfiglet R bindings for figlet; creates ascii logos. ghbutton
xkcdpw A package for making portmanteau-style passwords, as in this famous XKCD comic. Words are taken from scrapings of random Wikipedia pages. ghbutton
tfca Some methods for computing transfinite cardinal arithmetic. ghbutton
idgaf All public GitHub commits containing the “F-word” in the commit message, from 3/11/2012 to 7/24/2014. (warning: profanity) ghbutton



## Libraries

Package Description Download
libexpm High-performance C library for computing the matrix exponential of a dense matrix. ghbutton
meminfo A fully cross-platform C library for accessing basic hardware information, such as total RAM, free RAM, total swap, etc. Additionally, can get cache sizes and cache line sizes. Supports Linux, Windows, Mac OS X, and FreeBSD. Used by the R package “memuse” from the SHINRA project. ghbutton
ngram A C library for reading, processing, and “babbling” n-gram models. Used in the R package “ngram”. ghbutton



## Webapps

Large Projects

Package Description Download
TAG The Text Analytics Gateway (TAG) is an interactive webapp for performing simple analyses on unstructured text. It includes tools for importing data, preprocessing, and modeling. ghbutton
gmhelper A collection of utilities for game masters of fantasy-based role playing games, with a GUI and everything! ghbutton

Micro Projects

Package Description Download
hackR An interactive interface to the hackR package. Now you can hack the gibson without needing to learn R! ghbutton
livebutton
ngram An interactive interface to the ngram package. Ever wanted to make your very own wacky sentences using markov chains? Now you can! ghbutton
livebutton
RTop500 A visual summary of the performance over time (by various metrics) of the top 500 supercomputers since 1993. Written with R and shiny. Source is available in the RTop500 package. ghbutton
livebutton
Figlet as a Service A demo of the Rfiglet package. Written with R and shiny. Source is available in the Rfiglet package. ghbutton
livebutton