Two hundred and six new packages stuck to CRAN in March. Here are my “Top 40” selections in thirteen categories: Computational Methods, Data, Finance, Game Theory, Genomics, Machine Learning, Medicine, Networks, Science, Statistics, Time Series, Utilities, and Visualization.
ag5Tools v0.0.1: Offers tools for downloading and extracting data from the Copernicus Agrometeorological indicators from 1979 to present derived from reanalysis (AgERAS) dataset. See ag5Tools to get started and the vignette on extracting data.
AirMonitor v0.2.2: Provides utilities for working with hourly air quality monitoring data with a focus on small particulates (PM2.5) along with algorithms to calculate NowCast and the associated Air Quality Index (AQI) as defined by the US Environmental Projection Agency AirNow program. There is an Introduction, a Developers Style Guide, and a Data Model.
AFR v0.1.0: Provides tools for regression, prediction and forecast analysis of macroeconomic and credit data adapted for banking sector of Kazakhstan for bank analysts and non-statisticians. There are vignettes on Data transformatiom, Diagnostic Tests, and Regression.
fixedincome v0.0.1: Implements objects that abstract interest rates, compounding factors, day count rules, forward rates and term structure of interest rates to assist with calculations of interest rates and fixed income. Look here for examples.
socialranking v0.1.1: Offers a set of solutions to rank players based on a transitive ranking between coalitions, including CP-Majority, ordinal Banzhaf or lexicographic excellence solution summarized Allouche et al. (2020). See the vignette.
coda4microbiome v0.1.1: Provides tools for microbiome data analysis that take into account its compositional nature including functions for variable selection for both, cross-sectional and longitudinal studies, and for binary and continuous outcomes. See the vignette
ZetaSuite v1.0.0: Provides functions to score hits from two-dimensional RNAi screens analyze single cell transcriptomics to differentiate rare cells from damaged ones. See Vento-Tormo et al. (2018) for background and the vignette for examples.
sentiment.ai v0.1.1: Implements Sentiment Analysis via
tensorflow deep learning and gradient boosting models and also allows users to create embedding vectors for text which can be used in other analyses. See the vignette for an example.
sentopics v0.6.2: Offers a framework that joins topic modeling and sentiment analysis of textual data, and implements a fast Gibbs sampling estimation of Latent Dirichlet Allocation. See Griffiths & Steyvers (2004) and the Joint Sentiment/Topic Model of Lin, Everson & Ruger (2012). There is a vignette on Bascic Usage and another on Topical time series.
transforEmotion v0.1.0: Provides access to sentiment analysis using the
Python based huggingface transformer zero-shot classification model pipelines. The default pipeline is Cross-Encoder’s DistilRoBERTa trained on the Stanford Natural Language Inference and Multi-Genre Natural Language Inference datasets. There is a vignette on setting up
rts2 v0.3: Provides functions to support modelling case data for real-time surveillance of infectious diseases including functions to generate a computational grid over an area of interest and approximate log-Gaussian Cox Process model. See Diggle et al. (2013) and Solin and Särkkä (2020) for background and look here for examples.
BCDAG v1.0.0: Provides functions for structure learning of causal networks and estimation of joint causal effects from observational Gaussian data. See Castelletti & Mascaro (2021) and Castelletti & Mascaro (2022) for background, and the vignettes Random Data Generation, Output of
learn_DAG(), and MCMC scheme for posterior inference for details.
SEset v1.0.1: Implements tools to compute and analyze the set of statistically-equivalent Gaussian, linear path models which generate the input precision or (partial) correlation matrix. See README for examples.
bdc v1.1.0: Provides functions for biodiversity data cleaning organized into five themes: Merging datasets, Pre-filtering, Taxonomy, Space (Flagging low precision coordinates), and Time (flagging inconsistent data collection dates). There are vignettes on Standardization, Pre-filter, Space, Taxonomy, and Time.
cheem v0.2.0: Provides functions to explore local explanations of non-linear models by first calculating the tree SHapley Additive exPlanation for every observation and for calculating a projection basis, and then changing the basis with a radial tour. See Lundberg et al. (2019), Spyrison & Cook (2020) and Cook and Buja (2012) for background and the vignette to get started.
CondCopulas v0.1.2: Provides functions for the estimation of conditional copulas models, various estimators of conditional Kendall’s tau statistic as proposed in Derumigny and Fermanian 2019a, 2019b and 2020. See the vignette for examples.
multilevelmod v0.1.0: Implements bindings for hierarchical regression models for use with the
parsnip package. Models include longitudinal generalized linear models as described in Liang and Zeger (1986) and mixed-effect models as described in Pinheiro and Bates (2000). See the vignette for examples.
rbmi v1.1.3: Implements reference based multiple imputation allowing for the imputation of longitudinal datasets using predefined strategies. These include conventional MI methods, conditional mean imputation methods, and bootstrapped MI methods. See the Scope section of the Statistical Specifications vignette for more information on MI methods. There is a Quick Start Guide and an additional vignette on Advanced Functionality.
rlcv v1.0.0: Provides functions to estimate likelihood cross-validation bandwidth for uni- and multi-variate kernel densities which are robust with respect to fat-tailed distributions and outliers. See Wu (2019) for the theory and the vignette for an example.
workboots v0.1.1: Provides functions for generating bootstrap prediction intervals from a
tidymodels workflow. There is a Getting Started Guide and a vignette on estimating linear prediction intervals.
formatters v0.2.0: Provides a framework for rendering complex tables to ASCII, and a set of formatters for transforming values or sets of values into ASCII-ready display strings. See the vignette for examples.
ggmice v0.0.1: Provides functions to enhance a
mice imputation workflow with visualizations for incomplete and imputed data including functions to inspect missing data, develop imputation models, evaluate algorithmic convergence, and compare observed versus imputed data. See the Getting Started Guide and the vignette Old Firends.
langevitour v0.2: Implements an HTML widget that uses Langevin dynamics to show random walks through 2D projections of numerical data. It can be used from within R, or included in a self-contained Rmarkdown document. See the vignette for an examples.
picker v0.2.6: Provides functions to zoom, pan, and pick points from a
deck.gl scatterplot and includes tooltips, labels, a grid overlay, legends, and coupled interactions across multiple plots. See README for examples.