This package implements differentially private (DP) versions of common classical statistical procedures, including descriptive statistics (mean, variance, quantiles, histograms), hypothesis tests (t-test, chi-square, Kolmogorov-Smirnov, one-way ANOVA), and regression (closed-form DP linear regression and DP-SGD for generalized linear models). Provides Laplace and Gaussian mechanisms with analytic calibration, exponential mechanism for medians, privacy-aware confidence intervals that account for both sampling and privacy noise, and privacy budget accounting via basic, advanced, and Renyi differential privacy (RDP) composition. Designed for official statistics and privacy-preserving data analysis research.

The main functions in this package are with_cache() and cached_read(). The former is a simple way to cache an R object into a file on disk, using cachem'. The latter is a wrapper around any standard read function, but caches both the output and the file list info. If the input file list info hasn't changed, the cache is used; otherwise, the original files are re-read. This can save time if the original operation requires reading from many files, and/or involves lots of processing.

Citrus is a computational technique developed for the analysis of high dimensional cytometry data sets. This package extracts, statistically analyzes, and visualizes marker expression from citrus data. This code was used to generate data for Figures 3 and 4 in the forthcoming manuscript: Throm et al. â Identification of Enhanced Interferon-Gamma Signaling in Polyarticular Juvenile Idiopathic Arthritis with Mass Cytometryâ , JCI-Insight. For more information on Citrus, please see: Bruggner et al. (2014) <doi:10.1073/pnas.1408792111>. To download the citrus package, please see <https://github.com/nolanlab/citrus>.

Computes the minimum sample size required for the development of a new multivariable prediction model using the criteria proposed by Riley et al. (2018) <doi: 10.1002/sim.7992>. pmsampsize can be used to calculate the minimum sample size for the development of models with continuous, binary or survival (time-to-event) outcomes. Riley et al. (2018) <doi: 10.1002/sim.7992> lay out a series of criteria the sample size should meet. These aim to minimise the overfitting and to ensure precise estimation of key parameters in the prediction model.

This is an implementation of the algorithm described in Section 3 of Hosszejni and Frühwirth-Schnatter (2026) <doi:10.1016/j.jmva.2025.105536>. The algorithm is used to verify that the counting rule CR(r,1) holds for the sparsity pattern of the transpose of a factor loading matrix. As detailed in Section 2 of the same paper, if CR(r,1) holds, then the idiosyncratic variances are generically identified. If CR(r,1) does not hold, then we do not know whether the idiosyncratic variances are identified or not.

Infrastructure and functions that can be used for integrating Stan (Carpenter et al. (2017) <doi:10.18637/jss.v076.i01>) code into stand alone R packages which in turn use the CmdStan engine which is often accessed through CmdStanR'. Details given in Stan Development Team (2025) <https://mc-stan.org/cmdstanr/>. Using CmdStanR and pre-written Stan code can make package installation easy. Using staninside offers a way to cache user-compiled Stan models in user-specified directories reducing the need to recompile the same model multiple times.

Covered uses modern Ruby features to generate comprehensive coverage, including support for templates which are compiled into Ruby. It has the following features:

• Incremental coverage -- if you run your full test suite, and the run a subset, it will still report the correct coverage - so you can incrementally work on improving coverage.

• Integration with RSpec, Minitest, Travis & Coveralls - no need to configure anything - out of the box support for these platforms.

• It supports coverage of views -- templates compiled to Ruby code can be tracked for coverage reporting.

Rolling and expanding window approaches to assessing abundance based early warning signals, non-equilibrium resilience measures, and machine learning. See Dakos et al. (2012) <doi:10.1371/journal.pone.0041010>, Deb et al. (2022) <doi:10.1098/rsos.211475>, Drake and Griffen (2010) <doi:10.1038/nature09389>, Ushio et al. (2018) <doi:10.1038/nature25504> and Weinans et al. (2021) <doi:10.1038/s41598-021-87839-y> for methodological details. Graphical presentation of the outputs are also provided for clear and publishable figures. Visit the EWSmethods website for more information, and tutorials.

This package provides a pipeline for matching messy company name strings against a clean dictionary (e.g., Orbis'). Implements a cascading strategy: Exact -> Fuzzy ('zoomerjoin') -> FTS5 ('SQLite') -> Rarity Weighted. Name normalization covers German, French, Italian and English legal forms and conventions, which suits multilingual registers such as the Swiss one. Normalization discards detail, so several dictionary entries can collapse onto one string; these groups are matched once and a crosswalk back to every original entry is retained. References: Beniamino Green (2025) <https://github.com/beniaminogreen/zoomerjoin>; <https://www.sqlite.org/fts5.html>.

Calculates fundamental IO matrices (Leontief, Wassily W. (1951) <doi:10.1038/scientificamerican1051-15>); within period analysis via various rankings and coefficients (Sonis and Hewings (2006) <doi:10.1080/09535319200000013>, Blair and Miller (2009) <ISBN:978-0-521-73902-3>, Antras et al (2012) <doi:10.3386/w17819>, Hummels, Ishii, and Yi (2001) <doi:10.1016/S0022-1996(00)00093-3>); across period analysis with impact analysis (Dietzenbacher, van der Linden, and Steenge (2006) <doi:10.1080/09535319300000017>, Sonis, Hewings, and Guo (2006) <doi:10.1080/09535319600000002>); and a variety of table operators.

This is the very popular mine sweeper game! The game requires you to find out tiles that contain mines through clues from unmasking neighboring tiles. Each tile that does not contain a mine shows the number of mines in its adjacent tiles. If you unmask all tiles that do not contain mines, you win the game; if you unmask any tile that contains a mine, you lose the game. For further game instructions, please run `help(run_game)` and check details. This game runs in X11-compatible devices with `grDevices::x11()`.

Extended tools for analyzing telemetry data using generalized hidden Markov models. Features of momentuHMM (pronounced ``momentum'') include data pre-processing and visualization, fitting HMMs to location and auxiliary biotelemetry or environmental data, biased and correlated random walk movement models, hierarchical HMMs, multiple imputation for incorporating location measurement error and missing data, user-specified design matrices and constraints for covariate modelling of parameters, random effects, decoding of the state process, visualization of fitted models, model checking and selection, and simulation. See McClintock and Michelot (2018) <doi:10.1111/2041-210X.12995>.

This package contains a set of tools for constructing and coercing into and from the "mdate" class. This date class implements ISO 8601-2:2019(E) and allows regular dates and times to be annotated to express unspecified date or time components, approximate or uncertain components, ranges, and sets of dates. The package therefore retains, represents, and reasons about data and time imprecision, resolving to a single data/time only on demand. This is useful for describing and analysing temporal information, whether historical or recent, where date or time precision may vary.

This package implements Bayesian phase I repeated measurement design that accounts for multidimensional toxicity endpoints and longitudinal efficacy measure from multiple treatment cycles. The package provides flags to fit a variety of model-based phase I design, including 1 stage models with or without individualized dose modification, 3-stage models with or without individualized dose modification, etc. Functions are provided to recommend dosage selection based on the data collected in the available patient cohorts and to simulate trial characteristics given design parameters. Yin, Jun, et al. (2017) <doi:10.1002/sim.7134>.

Temporal disaggregation methods are used to disaggregate and interpolate a low frequency time series to a higher frequency series, where either the sum, the mean, the first or the last value of the resulting high frequency series is consistent with the low frequency series. Temporal disaggregation can be performed with or without one or more high frequency indicator series. Contains the methods of Chow-Lin, Santos-Silva-Cardoso, Fernandez, Litterman, Denton and Denton-Cholette, summarized in Sax and Steiner (2013) <doi:10.32614/RJ-2013-028>. Supports most R time series classes.

This package provides a package for summary and annotation of genomic intervals. Users can visualize and quantify genomic intervals over pre-defined functional regions, such as promoters, exons, introns, etc. The genomic intervals represent regions with a defined chromosome position, which may be associated with a score, such as aligned reads from HT-seq experiments, TF binding sites, methylation scores, etc. The package can use any tabular genomic feature data as long as it has minimal information on the locations of genomic intervals. In addition, it can use BAM or BigWig files as input.

Spatial homogeneous regions (SHRs) in tissues are domains that are homogenous with respect to cell type composition. We present a method for identifying SHRs using spatial transcriptomics data, and demonstrate that it is efficient and effective at finding SHRs for a wide variety of tissue types. concordex relies on analysis of k-nearest-neighbor (kNN) graphs. The tool is also useful for analysis of non-spatial transcriptomics data, and can elucidate the extent of concordance between partitions of cells derived from clustering algorithms, and transcriptomic similarity as represented in kNN graphs.

Calculates distances from point locations to features. The usual approach for eg. resource selection function analyses is to generate a complete distance to features surface then sample it with your observed and random points. Since these raster based approaches can be pretty costly with large areas, and often lead to memory issues in R, the distanceto package opts to compute these distances using efficient, vector based approaches. As a helper, there's a decidedly low-res raster based approach for visually inspecting your region's distance surface. But the workhorse is distance_to.

This package provides a Bayesian model selection approach for generalized linear mixed models. Currently, GLMMselect can be used for Poisson GLMM and Bernoulli GLMM. GLMMselect can select fixed effects and random effects simultaneously. Covariance structures for the random effects are a product of a unknown scalar and a known semi-positive definite matrix. GLMMselect can be widely used in areas such as longitudinal studies, genome-wide association studies, and spatial statistics. GLMMselect is based on Xu, Ferreira, Porter, and Franck (202X), Bayesian Model Selection Method for Generalized Linear Mixed Models, Biometrics, under review.

Identifies regime changes in streamflow runoff not explained by variations in precipitation. The package builds a flexible set of Hidden Markov Models of annual, seasonal or monthly streamflow runoff with precipitation as a predictor. Suites of models can be built for a single site, ranging from one to three states and each with differing combinations of error models and auto-correlation terms. The most parsimonious model is easily identified by AIC, and useful for understanding catchment drought non-recovery: Peterson TJ, Saft M, Peel MC & John A (2021) <doi:10.1126/science.abd5085>.

There are occasions where you need a piece of HTML with integrated styles. A prime example of this is HTML email. This transformation involves moving the CSS and associated formatting instructions from the style block in the head of your document into the body of the HTML. Many prominent email clients require integrated styles in HTML email; otherwise a received HTML email will be displayed without any styling. This package will quickly and precisely perform these CSS transformations when given HTML text and it does so by using the JavaScript juice library.

Allows for fitting of maximum likelihood models using Markov chains on phylogenetic trees for analysis of discrete character data. Examples of such discrete character data include restriction sites, gene family presence/absence, intron presence/absence, and gene family size data. Hypothesis-driven user- specified substitution rate matrices can be estimated. Allows for biologically realistic models combining constrained substitution rate matrices, site rate variation, site partitioning, branch-specific rates, allowing for non-stationary prior root probabilities, correcting for sampling bias, etc. See Dang and Golding (2016) <doi:10.1093/bioinformatics/btv541> for more details.

Computes A-, MV-, D- and E-optimal or near-optimal block designs for two-colour cDNA microarray experiments using the linear fixed effects and mixed effects models where the interest is in a comparison of all possible elementary treatment contrasts. The algorithms used in this package are based on the treatment exchange and array exchange algorithms of Debusho, Gemechu and Haines (2018) <doi:10.1080/03610918.2018.1429617>. The package also provides an optional method of using the graphical user interface (GUI) R package tcltk to ensure that it is user friendly.

Server-side plumbing for the ui.tsx pattern in Shiny': the user interface is defined in a client React (<https://react.dev/>) bundle, and the Shiny server contains only reactive computation. Provides page builders that discover and serve the client bundle, a render function that publishes any JSON-serializable value to the client, and custom messages to React components. Ships no user interface components, so the app author owns the whole front end. The React runtime and the client hooks are bundled, so no JavaScript build step is required to get started.