Changes in version 0.1.6                        

  - Added pointpars argument to pirateplot() that allows the user to
    specify plotting parameters for individual points (e.g.; colors,
    labels, pch). See the pirateplot vignette for details.

  - Added yaxt.y argument to pirateplot() that allows you to specify
    where yaxis tick marks are located.

  - Added beside argument to pirateplot(). When beside = TRUE values of
    second independent variables are shown in separate plots rather than
    beside each other.

  - You can now enter data as a wide dataframe or matrix in pirateplot()
    without specifying a formula. Each column will be shown in a
    different bean.

                 Changes in version 0.1.5 (2017-04-19)                  

  - Minor changes to theme = 1 in pirateplot(). Changed default value of
    cap.beans to TRUE

  - Added color mixing arguments mix.col and mix.p to piratepal(). These
    allow you to mix the default palettes with a specified color (e.g.;
    "white")

  - Added the option to specify data in pirateplot() as a list of
    numeric vectors, or as a numeric dataframe or matrix without
    specifying a formula. Each column / element will be taken as a new
    group.

  - New palettes in piratepal(): decision.

  - Fixed bug in sortx in pirateplot(). Sorting data by functions (e.g.
    sortx = "mean") should now work.

  - Added gl argument to pirateplot() to specify locations of gridlines
    (e.g.; gl = seq(0, 10, 1))

  - Added cex.names argument to control size of bean names (currently
    this was controlled by cex.lab, which now controls the size of the
    axis names.)

                 Changes in version 0.1.4 (2017-02-06)                  

  - Some minor changes to default plotting parameters that I think make
    the default plots look a bit nicer.

  - Added cap.beans argument to pirateplot(). When cap.beans = TRUE,
    beans will be cut at the maximum and minimum values of the data.

  - Added cap.beans argument to pirateplot(). When cap.beans = TRUE,
    beans will be cut at the maximum and minimum values of the data.

  - Added two new inf.method values: sd for standard deviation, and se
    for standard error

                 Changes in version 0.1.2 (2016-10-25)                  

  - Minor updates to themes. Added theme = 3

  - You can now assign a pirateplot to a variable to return summary
    statistics.

                        Changes in version 0.1.0                        

  - Added a NEWS.md file to track changes to the package.

  - Re-structured code generating colors and opacities in pirateplot()
    to make future updates easier.

  - Added quant, quant.length and quant.width arguments that add
    horizontal lines for specified quantiles to each bean (thanks
    @pat-s)

  - Added several new arguments (e.g.; bean.fill.col for customising
    pirateplots

  - Improved theme support (now under theme rather than theme.o)

                        Changes in version 0.0.7                        

  - pirateplot() can now handle up to 3 IVs!
    Example: pirateplot(age ~ sex + headband + favorite.pirate, data =
    pirates).
    Levels of the third IV are shown in separate plots in a grid.

Minor and Bug-fixes

  - The inf.p parameter in pirateplot() was prevously not being passed
    to the Bayesian HDIs, rendering all inference bands to be the
    default of 95% (thanks to Roman Pahl for catching this). This has
    now been fixed.

                        Changes in version 0.0.6                        

  - Added hdi.band argument to pirateplot(). Setting hdi.band = "tight"
    will constrain inference bands to bean densities.
  - Minor changes to gridlines when specifying gl.col.