Linear and Smooth Predictor Modelling with Penalisation and Variable Selection

Fit a model with potentially many linear and smooth predictors. Interaction effects can also be quantified. Variable selection is done using penalisation. For l1-type penalties we use iterative steps alternating between using linear predictors (lasso) and smooth predictors (generalised additive model).

Intensity Analysis of Spatial Point Patterns on Complex Networks

Tools to analyze point patterns in space occurring over planar network structures derived from graph-related intensity measures for undirected, directed, and mixed networks. This package is based on the following research: Eckardt and Mateu (2018) . Eckardt and Mateu (2021) .

Aggregate Numeric, Date and Categorical Variables

Convenience functions for aggregating a data frame or data table. Currently mean, sum and variance are supported. For Date variables, the recency and duration are supported. There is also support for dummy variables in predictive contexts. Code has been completely re-written in data.table for computational speed.

Scalable Gaussian-Process Approximations

Fast scalable Gaussian process approximations, particularly well suited to spatial (aerial, remote-sensed) and environmental data, described in more detail in Katzfuss and Guinness (2017) . Package also contains a fast implementation of the incomplete Cholesky decomposition (IC0), based on Schaefer et al. (2019) and MaxMin ordering proposed in Guinness (2018) .

Multiple Primary Endpoints

Functions for calculating sample size and power for clinical trials with multiple (co-)primary endpoints.

Estimation in Adaptive Group Sequential Trials

Calculation of repeated confidence intervals as well as confidence intervals based on the stage-wise ordering in group sequential designs and adaptive group sequential designs. For adaptive group sequential designs the confidence intervals are based on the conditional rejection probability principle. Currently the procedures do not support the use of futility boundaries or more than one adaptive interim analysis.

Functional Rarity Indices Computation

Computes functional rarity indices as proposed by Violle et al. (2017) . Various indices can be computed using both regional and local information. Functional Rarity combines both the functional aspect of rarity as well as the extent aspect of rarity. 'funrar' is presented in Grenié et al. (2017) .

Quaternions Splines

Provides routines to create some quaternions splines: Barry-Goldman algorithm, De Casteljau algorithm, and Kochanek-Bartels algorithm. The implementations are based on the Python library 'splines'. Quaternions splines allow to construct spherical curves. References: Barry and Goldman , Kochanek and Bartels .

A Time Series Toolbox for Official Statistics

Plot official statistics' time series conveniently: automatic legends, highlight windows, stacked bar chars with positive and negative contributions, sum-as-line option, two y-axes with automatic horizontal grids that fit both axes and other popular chart types. 'tstools' comes with a plethora of defaults to let you plot without setting an abundance of parameters first, but gives you the flexibility to tweak the defaults. In addition to charts, 'tstools' provides a super fast, 'data.table' backed time series I/O that allows the user to export / import long format, wide format and transposed wide format data to various file types.

Create Waffle Chart Visualizations

Square pie charts (a.k.a. waffle charts) can be used to communicate parts of a whole for categorical quantities. To emulate the percentage view of a pie chart, a 10x10 grid should be used with each square representing 1% of the total. Modern uses of waffle charts do not necessarily adhere to this rule and can be created with a grid of any rectangular shape. Best practices suggest keeping the number of categories small, just as should be done when creating pie charts. Tools are provided to create waffle charts as well as stitch them together, and to use glyphs for making isotype pictograms.