riekelib 0.3.0

riekelib 0.3.0

New features

• Additional covariance functions for gaussian processes:
  • cov_rational()
  • cov_periodic()
  • cov_linear()
  • cov_noise()

Conscious decoupling

Per #22, I'm moving functions that are specific for manipulating Qualtrics survey data for Memorial Hermann into their own package, {heRmann}.

The following functions from riekelib 0.2.0 have been moved:

• fetch_surveys()
• deduplicate_ip()
• fix_survey_names()
• rnps()
• qnps()

The following functions were available in riekelib 0.2.0.9000 as of 8/22/23, but were removed on 8/23/23.

• set_recode_path()
• read_recodes()
• fix_campus_names()

riekelib 0.2.0

riekelib 0.2.0

New features

Gaussian process functions

• cholesky_decompose(): generate the lower triangular Cholesky decomposition of
  a matrix (consistent with Stan's implementation).
• cov_exp_quad(): generate a covariance matrix using the exponentiated
  quadratic kernel.
• condition_gaussian_process(): generate new draws from a multivariate normal
  distribution by conditioning on a Gaussian Process.

ggplot2 functions

• scale_xy_percent(), scale_x_percent(), and scale_y_percent() provide
  shortcuts for formatting axis labels with scales::label_percent().
• scale_xy_comma(), scale_x_comma(), and scale_y_comma() provide shortcuts
  for formatting axis labels with scales::label_comma().

Minor bug fixes and improvements

• ggquicksave() now uses ragg by default.
• theme_rieke() now uses the fonts Tiempos Text and IBM Plex Sans for the
  title and other text, respectively, by default.
• Where applicable, functions now use the .data pronoun from
  rlang. This removes the need for globals.R.

riekelib 0.1.0

riekelib 0.1.0

A collection of functions I use regularly:

• Math and stats extensions
  • logit(), expit()
  • softmax()
  • beta_interval(), normal_interval()
• Tidyverse extensions
  • arrange_vector()
  • percent()
  • theme_rieke()
  • ggquicksave()
  • color_text()
• Qualtrics extensions
  • fetch_surveys()
  • fix_survey_names()
  • deduplicate_ip()
  • rnps(), qnps()