Notable New and Updated R packages (to October 2016)

As we prepare for the upcoming release of Microsoft R Open, I've been preparing the list of new and updated packages for the spotlights page. This involves scanning the CRANberries feed (with gracious thanks to Dirk Eddelbuettel) for newly-released packages and significant updates to existing ones.

This is a lot of data to process. For context, in October 2016 alone there were 174 new packages and 629 updated packages. For the 4 months since the last MRO release, I reviewed all of the new packages and some of the updates (those with a 1.0 or 2.0 release) by hand, and selected those that are likely to be of interest to a broad segment of R users. I hope you find some useful nuggets in the list. It's likely I missed some good ones; if you know of any others that should have been included, let me know in the comments.

Notable New and Updated R Packages: July — October 2016

Data Munging

anytime: Convert anything that looks vaguely like a date into a POSIXct object. Examples here.

fasttime: Fast functions for timestamp manipulation.

forcats: Tools for working with categorical variables (factors).

roll: Parallel functions for computing rolling statistics of time-series data.

stringb: String handling functions with an interface like stringr/stringi, but without the dependencies.

tidyverse: A collection of popular packages for data munging from Hadley Wickham. More here.

 

Data Sources

countyweather: Meterological time series data for U.S. counties, from NOAA.

darksky: Interface to the Dark Sky API, which provides current or historical global weather conditions.

energyr: Data published by the United States Federal Energy Regulatory Commission.

genderNames: API client for genderize.io which will tell you the gender of the name you input.

Rlinkedin (updated): Provides access to the LinkedIn API for people, job, and discussion data.

outbreaks: A collection of disease outbreak data, some real, some simulated.

rnoaa: Download NOAA weather data. Vignettes provided for many different data types.

rpinterest: Access the Pinterest API for data on boards, pins and users.

ttbbeer: Beer statistics from the U.S. Department of the Treasury, Alcohol and Tobacco Tax and Trade Bureau.

ubeR: Interface to the Uber API.

 

Graphics

cartogram: Create and plot a continuous area cartogram by a rubber sheet distortion algorithm.

colourlovers: Provides access to the COLOURlovers API, which offers color inspiration and color palettes.

earthtones: Derive a color palette from a satellite image of a location on Earth. The vignette includes examples.

Homeric: Doughnut plots - pie charts with a blank center. The package is named after Homer Simpson.

RSVGTipsDevice (updated): An R SVG graphics device with support for dynamic tips and hyperlinks.

tweenr (updated): Create smooth animations by interpolating data.

 

Interfaces

haven (updated): Import and export SPSS, Stata and SAS Files. Details here.

readr (updated): Read flat/tabular text files from disk (or a connection). Overview here.

SnakeCharmR: Run Python code, make function calls, and assign and retrieve variables from R.

sofa: Provides an interface to the NoSQL database CouchDB.

sparklyr: Provision, connect and interface to Apache Spark from within R. Details at spark.rstudio.com.

XR: Support for interfaces from R to other languages, described in the book “Extending R” by John Chambers.

 

Programming Tools

checkpoint (updated): Easily use R packages from a specific date for reproducibility (see the vignette). This update adds the ability to delete snapshots and use a local snapshot repository.  

curl (updated): A modern and flexible web client for R. Vignette included.

curry: Partial function application with %<%, %-<%, and %><%.

jsonlite (updated): fast JSON parser and generator optimized for statistical data and the web.

proto (updated): An object oriented system using object-based rather than class-based object oriented ideas, maintained by Hadley Wickham. Includes a vignette.

Rfast: Fast covariance matrix calculation, Mahalanobis distance and column-wise variances and other fast reimplementations of R functions.

sticky: Allows objects to be marked as 'sticky' and have resilient attributes that persist during many operations.

types: Simple type annotation for R, as a convention to allow other packages to provide error checking, automatic documentation or optimizations.

 

Reporting

bookdown: Output formats and utilities for authoring books with R Markdown.

htmltidy: Clean up gnarly HTML generated by the likes of Google Docs and Microsoft Word.

rmarkdown (updated): Dynamic documents for R. Overview and examples here.

tables (updated): Compute and display complex tables of summary statistics. Examples in the vignette.

tint: Format RMarkdown documents in the style of an Edward Tufte book. Examples here.

 

Statistical Modeling and Analysis

arc: Classification-based on Association Rules (CBA) algorithm for association rule classification (ARC).

brms: Bayesian regression models using Stan. Vignette included.

CADStat: Provides a GUI to several statistical methods including generalized linear regression, conditional probability calculations, and regression trees. The manual provides more details.

endogenous: Classical Simultaneous Equation Models.

ensembleR: Functions to use ensembles of several machine learning models specified in caret package.

MSGARCH: Markov-Switching GARCH Models.

oddsratio: Odds Ratio Calculation for Generalized Additive Models and Generalized Linear Models.

pkmon: Least-squares estimators under k-monotony constraint. Pikachu not included.

radiant: Business analytics using R and Shiny. Overview here.

simputation: Missing value imputation methods that work with the magrittr %>% pipe operator. Vignette included.

spikes: Detect election fraud from irregularities in vote-share distributions.

SurvDisc: Functions for discrete time survival analysis and longitudinal analysis, from the FDA.

 

Utilities

convertr: Convert between a broad range of scientific, historical, and industrial unit types.

digitize: Extract data from images of printed charts.

measurements: Convert between metric and imperial measurement units.

qrencoder: Create scannable QR codes and matrix barcodes.

optimr: A replacement and extension of the 'optim' function for smooth, possibly box constrained functions of several or many parameters.

RDocumentation: Display help files as they appear on RDocumentation.org

triangulation: Determine the location of an observer based on the angle towards three known points.

 

Other

FuzzyR: Design and simulate fuzzy logic systems using Type 1 Fuzzy Logic.

statquotes: Random quotations on topics in statistics, data visualization and science.

wfindr: Crossword, Scrabble and anagram solver.

 

You can find more notable packages from prior releases of MRO at the Spotlights page on MRAN.

Comments

[vonjd]

I also published a new release of the OneR machine learning package: https://CRAN.R-project.org/package=OneR

You can find the vignette here to get you started: https://cran.r-project.org/web/packages/OneR/vignettes/OneR.html

Additionally I made a quick-start video: https://www.youtube.com/watch?v=AGC0oRlXxgU

[Ian]

Errorizer is brand new, although I guess it came out in November, so maybe next round.

https://CRAN.R-project.org/package=errorizer

[Marcin Kosiński]

recordr is no longer on CRAN
https://cran.r-project.org/web/packages/recordr/index.html

[David Smith]

Thanks Marcin, I've removed recordr from the list. I wonder what the policy violation was.

[WinVectorLLC]

Fun list. I’d like to suggest researchers with “data munging” needs also check out our package “vtreat” https://mran.microsoft.com/package/vtreat/