Changes in Version 2.3

RELEASE DATE: 1 October 2015

MINOR CHANGES

• new argument added for dotplots to allow group labels to be placed inside the plot, rather than in the axis margin: internal.labels=TRUE

PATCH 2.3.1 - 06/10/2015

• fix how viewports are named

PATCH 2.3.2 - 08/10/2015

• fix a bug that prevented comparison lines to be drawn on categorical dotplots

PATCH 2.3.3 - 13/10/2015

• fix a bug where adding comparison intervals would break the plot if any of the subgroups in a dotplot were too small. New behaviour ignores small groups and only compares large ones.

PATCH 2.3.4 - 13/10/2015

• continuing from 2.3.3, but now use independent covariances

PATCH 2.3.4 - 13/10/2015

• FIX for Lite: dotplots by factor labels now also on histograms

PATCH 2.3.5 - 28/10/2015

• fix the "number of missing observations" shown in summary output

PATCH 2.3.7 - 02/11/2015

• no longer redraws dotplots; instead, passes a logical attribute for whether or not the scaling has changed (and therefore that the plot should be redrawn).

PATCH 2.3.8 - 16/11/2015

• fix bug in colby if the variable has only one unique value

Changes in Version 2.2

RELEASE DATE: 14 September 2015

NEW FEATURES - SURVEY DESIGN

• confidence intervals for histograms and bar plots

• comparison intervals for histograms broken down by a factor

• summary information for all basic plots (histograms, bar plots, and scatter plots)

BUG FIXES

• fixes a bug where missing information on barplots and scatter plots would cause the plotting function to die

• fixes a bug in the printing of summary objects

• fixes a bug where the minimum value of a single numeric variable summary was ommited

• and various other small bug fixes

Changes in Version 2.1

RELEASE DATE: 04 August 2015

NEW FEATURES

• allow zooming of plots with the new zoom argument. Works for both univariate and bivariate plots, and a related funcionality for 'zooming in' on bars in a barplot.

PATCH 2.1-1: 28 August 2015

• fix a bug that occurs when all survey weights are equal

Changes in Version 2.0.6

RELEASE DATE: 03 August 2015

BUG FIXES

• fix an issue where requesting summary of 'dotplots' resulted in creating a new device, which resulted in errors on the Shiny server.

Changes in Version 2.0.5

RELEASE DATE: 27 July 2015

NEW FEATURES

• additional arguments 'xlim' and 'ylim' allow users to specify the range of values shown on the plot

BUG FIXES

• several issues for dotplots have been fixed

• weighting variable used when drawing a scatter plot of survey data

• conf now corresponds to Year 12 intervals in dot plot inferences (previously, comp corresponded to this interval)

Changes in Version 2.0.3

RELEASE DATE: 01 July 2015

BUG FIXES

• fix up the order of bars in segmented bar plots to correspond to the legend

Changes in Version 2.0.2

RELEASE DATE: 24 June 2015

MINOR CHANGES

• remove facility where the colour-by variable is ignored if there are 'too many' levels—this is now left up to users to decide if colour by a particular variable makes sense of not.

Changes in Version 2.0.1

RELEASE DATE: 16 June 2015

MINOR CHANGES

• Dotplot locating implemented using new methodology, with the additional argument 'label.extreme = numeric(2)', allowing users to specify how many lower and upper points to identify, respectively.

• equivalently, extreme points (by using Mahalanobis' distance) can be labelled on scatter plots using 'label.extreme = numeric(1)'.

BUG FIXES

• fixed a small bug that stopped inference from working in dotplots when 'x' is a factor and 'y' is the numeric variable.

• fix a bug that caused 'nbins = 0' in some cases.

Changes in Version 2.0

Release Date: 26 May 2015

MAJOR CHANGES

• The entire package has been rewritten to accomodate complex survey designs. At present, survey objects are not fully supported, however the functionality will be added over time.

• A huge reduction in computation requirements for plots to increase efficiency.

• Algorithms used to compute inference intervals have been modified to use iNZightMR for comparisons.

• lots of other changes to layout and presentation

• added additional arguments locate, locate.id, locate.col (and others) for locating points by IDs. This is used in the improved locator functinoality in the main iNZight program.

MINOR CHANGES

• the col.by and size.by arguments have been replaced by colby and sizeby

• documentation has been added for several of the functions (finally!)

Changes in Version 1.0.3

BUG FIXES

• specifying ‘g2.level' with numbers wasn’t working, has been fixed for plots, summary and inference information.

• added more space to the y-axis on scatter plots

Changes in Version 1.0.2

BUG FIXES

• An error where the response was printed instead of the x-variable name in summary output for quadratic curves has been fixed.

Changes in Version 1.0.1

MINOR CHANGES

• The type of plot used can be specified by setting the 'largesample' argument. When set to 'TRUE', it uses the histogram or grid-density plot; when 'FALSE' it uses the dotplot or scatter plot. If set to 'NULL', it uses the sample size to determine which plot to draw (default).

• To allow identification features and any additional features to be added to plots afterwards, the last viewport is the one surrounding the main plot (excluding the plot labels and legend). Note that this only works if the data haven't been broken down by 'g1'.

• Display which variables cannot be plot due to too many levels, as well as the number of levels, when attempting to draw bar plot. (max levels = 101).

• Trend lines and smoothers added to the legend.

• Alternative method of shading grid-tiles on the grid-density plot using quantiles rather than absolute counts. This prevents large counts having too large of an influence.

BUG FIXES

• A bug where the grid-density plot is not using the correct scale has been fixed.

MAJOR RELEASE VERSION 1.0.0

• New major release of iNZightPlots released, completely rewritten using 'grid'.