Last month, 217 new packages were submitted to CRAN. By my count this was the highest monthly value recorded for the past nine years. Below, are brief descriptions of fifty-two of these new-for-December packages grouped into six categories: Data, Data Science, Financial Analysis, Statistics, Utilities and Visualizations. The Financial Analysis category is noteworthy. It is unusual to have six interesting packages centered on some financial application in a given month.
It is also unusual for me to be able to identify a favorite package. However, this month I am comfortable pointing out deconvolveR from Bradley Efron and Balasubramanian Narasimhan which implements Empirical Bayes methods for learning prior distributions from data. I expect that this package will become important to many Bayesian and “Almost Bayesian” statisticians.
BETS v0.0.98: Provides access to Brazilian economic time series from the Getulio Vargas Foundation, the Central Bank of Brazil and the Brazilian Institute of Geography and Statistics.
lingtypology v1.0.0: Allows users to access Glottolog database of world languages. The vignette contains several examples including how to use the map.feature function to produce maps similar to those used in Cross-Linguistic Linked Data projects.
pipeliner v0.1.1: Defines a framework for defining ‘pipelines’ of functions for data transformations, model estimation and inverse-transformations that results model-scoring functions that automatically apply the entire pipeline. The vignette explains the concept and provides examples.
sqlscore v0.1.0: Provides utilities for generating SQL queries from R model objects. It generates the SQL to score a generalized linear model. See README.
FinancialMath v0.1.1: Provides function to help prepare for the Society of Actuaries and Casualty Actuarial Society ‘Financial Mathematics’ exam.
tidyquant v0.1.0: Bringing quantitative financial analysis to the ‘tidyverse’ this package provides wrappers for various xts, quantmod and TTR package functions and returns the objects in the tidy tibble format. The vignette provides an introduction.
valuer v1.1.0: Implementing the pricing framework described in Bacinello et al. and the fee structure discussed in Bernard et al., this package provides to price variable annuity life insurance contracts by means of Monte Carlo methods. See the vignette for an introduction.
arsenal v0.1.2: Provides functions for large-scale statistical summaries which take advantage of work RStudio reporting tools. There are vignettes describing the freqlist, modelsum, and tableby functions.
deconvolveR v1.0-3: Provides Empirical Bayes methods for learning prior distributions from data using deconvolution and Empirical Bayes methods. The vignette on Empirical Bayes Deconvolution by Balasubramanian Narasimhan and Bradley Efron is really good reading.
dynamichazard v0.1.0: Contains functions for fitting dynamic hazard models with binary outcomes using state space models using methods described in Fahrmeir (1992) and Fahrmeir (1994). There are vignettes on Comparing methods for time varying logistic models and the theory behind the ddhazard function.
expint v0.1-2: Provides functions that implement exponential integrals and the incomplete gamma function defined for negative values of its first argument. The vignette describes the math, the underlying C routines and alternative methods of calculation.
joinRML v0.1.1: Provides functions to fit joint models of longitudinal data and time to event outcomes as proposed by Henderson et. al., but extended to the case of multiple continuous longitudinal measures. There is an introduction to the package and a vignette with the technical details.
blob v1.0.0: Provides an S3 class for representing vectors of binary data.
replyr v0.2.0: Provides methods for working with remote ‘tbl’ sources (‘SQL’ databases, ‘Spark’) through ‘dplyr’. There is an introduction and vignettes for NA removal, parametric programming, and using the let function.