By my count, 157 new packages stuck to CRAN in December. Below are my “Top 40” picks in ten categories: Computational Methods, Data, Finance, Machine Learning, Medicine, Science, Statistics, Time Series, Utilities and Visualization. This is the first time I have used the Medicine category. I am pleased that a few packages that appear to have clinical use made the cut. Also noteworthy in this month’s selection are the inclusion of four packages from the Microsoft Azure team (stuffing 41 packages into the “Top 40”), and some eclectic, but fascinating packages in the Science section.
rdhs v0.6.1: Implements a client querying the DHS API to download and manipulate survey datasets and metadata. There are introductions to using rdhs and the rdhs client, an extended example about Anemia prevalence, and vignettes on Country Codes, Interacting with the geojson API results, and Testing.
riskParityPortfolio v0.1.1: Provides functions to design risk parity portfolios for financial investment. In addition to the vanilla formulation, where the risk contributions are perfectly equalized, many other formulations are considered that allow for box constraints and short selling. The package is based on the papers: Feng and Palomar (2015), Spinu (2013), and Griveau-Billion et al.(2013). See the vignette for an example.
BTM v0.2: Provides functions to find
Biterm topics in collections of short texts. In contrast to topic models, which analyze word-document co-occurrence, biterms consist of two words co-occurring in the same short text window.
ParBayesianOptimization v0.0.1: Provides a framework for optimizing Bayesian hyperparameters according to the methods described in Snoek et al. (2012). There are vignettes on standard and advanced features.
metaRMST v1.0.0: Provides functions that use individual patient-level data to produce a multivariate meta-analysis of randomized controlled trials with the difference in restricted mean survival times ( RMSTD ).
tabula v1.0.0: Provides functions to examine archaeological count data and includes several measures of diversity. There are vignettes on Diversity Measures, Matrix Classes, and Matrix Seriation. This last vignette includes an example reproducing the results of Peeples and Schachner (2012).
areal v0.1.2: Implements areal weighted interpolation with support for multiple variables in a workflow that is compatible with the
sf frameworks. There are vignettes on Areal Interpolation, Wieghted Areal Interpoaltion, and Preparing Data for Interpolation.
mulitRDPG v1.0.1: Provides functions to fit the Multiple Random Dot Product Graph Model and performs a test for whether two networks come from the same distribution. See Nielsen and Witten (2018) for details.
ocp v0.1.0: Implements the Bayesian online changepoint detection method of Adams and MacKay (2007) for univariate or multivariate data. Gaussian and Poisson probability models are implemented. The vignette provides an introduction.
smurf v1.0.0: Implements the SMuRF algorithm of Devriendt et al. (2018) to fit generalized linear models (GLMs) with multiple types of predictors via regularized maximum likelihood. See the package Introduction.
subtee v0.3-4: Provides functions for naive and adjusted treatment effect estimation for subgroups. Proposes model averaging Bornkamp et al. (2016) and bagging Rosenkranz (2016) to address the problem of selection bias in treatment effect estimation for subgroups. There is a Introduction and vignettes for the plot and subbuild functions.
xspliner v0.0.2: Provides functions to assist model building using surrogate black-box models to train interpretable spline based, additive models. There are vignettes on Basic Theory and Usage, Automation, Classification, Use Cases, Graphics, Extra Information, and the xspliner Environment.
mfbvar v0.4.0: Provides functions for estimating mixed-frequency Bayesian vector autoregressive (VAR) models with Minnesota or steady-state priors as those used by Schorfheide and Song (2015), or by Ankargren, Unosson and Yang (2018). Look at the GitHub page for an example.
NTS v1.0.0: Provides functions to simulate, estimate, predict, and identify models for nonlinear time series.
AzureContainers v1.0.0: Implements an interface to container functionality in Microsoft’s
Azure cloud that enables users to manage the the
Azure Container Instance,
Azure Container Registry, and
Azure Kubernetes Service. There are vignettes on Plumber model deployment and Machine Learning server model deployment.
AzureRMR v1.0.0: Implements lightweight interface to the Azure Resource Manager REST API. The package exposes classes and methods for
OAuth authentication and working with subscriptions and resource group. There is an Introduction and a vignette on Extending AzureRMR.
cliapp v0.1.0: Provides functions that facilitate creating rich command line applications with colors, headings, lists, alerts, progress bars, and custom CSS-based themes. See the README for examples.
remedy v0.1.0: Implements an RStudio Addin offering shortcuts for writing in