Both installed data sets, vaccinations and majors, are better documented.
The a field of vaccinations (the within-survey fraction of respondents, which can be computed from the other fields) has been removed, and the start_date and end_date fields (Dates, obtained from the ALP website) have been added.
The following changes broke no examples or tests but could change behavior in rare cases:
is_lodes_form() now returns FALSE if any axis-alluvium pairs are duplicated, and throws the previous warning as a message. This should be more helpful than the previous behavior of suppressing the warning and leaving tidyr::gather() to throw an error referring to rows of the already-transformed internal data.stratum = NULL and alluvium = NULL have been added to the stats. This prevents the “unknown aesthetics” warnings that print when these aesthetics are passed to layers rather than to the plot initialization.lode.orderingFor consistency with the behavior of aes.bind, stat_alluvium() now invokes lode.ordering together with lode.guidance: If the vectors of lode.ordering include duplicates, i.e. they do not completely determine an order, then the remaining deposits are used to refine the order. Previously, lode.ordering was assumed to consist of permutation vectors, so the two parameters were mutually exclusive.
Additionally, for consistency with other influences on the lode order, the vectors of lode.ordering are reversed if reverse = TRUE (the default). This will change some plots but will not produce new errors.
The lode.ordering parameter of stat_alluvium() has been deprecated. Instead, the new order aesthetic gives priority to its argument over the differentiation aesthetics in arranging the lodes within each stratum, without producing graphical artifacts. This aesthetic can also be used in stat_flow().
Alluvia within “deposits” are now consistently ordered in positive and negative strata, rather than according to absolute. This avoids the “twisting” of flows between strata of different signs. Whereas the orderings of the deposits matter to the stacked-histogram reading of the plot, the orderings of the alluvia should simply maximize its elegance and readability. This will change some plots but will not produce new errors.
The alluvial stats now compute four variables for use with after_stat(): numeric variables n, count, and prop; and character variables lode (when the alluvium aesthetic is specified) and flow (when using the flow stat). The numerical variables can be weighted using the weight aesthetic, which is dropped during computation (so that it does not confuse the geoms), while lode is distilled according to a new distill parameter. This use of weight may cause confusion with its use by the is_*_form() functions until they are upgraded in the next version.
These new variables complement the already-computed but heretofore undocumented variables stratum and deposit. stratum obviates the need for the infer.label parameter, which is deprecated. Its alias, label.strata, is now defunct. (The variable alluvium is often computed, but it is manipulated to be used by the geom layers and should not be passed to an aesthetic.) deposit takes contiguous integer values forward along the axes and upward along the (signed) strata at each axis.
The knot.pos parameter of geom_alluvium() and geom_flow() is now interpreted as a proportion of the total length of each flow, i.e. of the gap between adjacent strata (not axes). This means that values will vary with axis positions and stratum widths. Setting the new knot.prop parameter to FALSE prevents this by interpreting knot.pos as a constant value in the x direction.
These flows are rendered using grid::xsplineGrob() with four control points each: the endpoints and the two knots. To complement them, several other curves are now available: linear (equivalent to knot.pos = 0), cubic, quintic, sinusoidal, arctangent, and sigmoid, summoned by the new curve_type parameter (which defaults to the x-spline). (The asymptotic functions, arctangent and sigmoid, are compressed according to the new curve_range parameter.) The new curves are rendered piecewise linearly, with resolution controlled by the new segments parameter (similar to ggplot2::stat_ellipse()).
The stratum and lode ordering parameters now default to NULL, in which case they are reassigned to global options internally. This simplifies their documentation. The new curve parameters curve_type, curve_range, and segments can also be set as options and are documented in the same way.
In response to ggplot2 v3.2.0, which removes the plyr dependency, the dependency has been removed from ggalluvial as well.
The function self_adjoin() is debugged for use with a continuous-valued x variable. An example, taking x to be the date of each vaccination survey in vaccinations, is documented with stat_stratum().
This patch fixes a bug with including negative observations in alluvia-form data due to outdated code that prohibited negative y values. This was discovered while drafting two examples of this usage, which are included in the documentation.
min.height and max.height parameters of stat_stratum() are deprecated in favor of min.y and max.y (which better adhere to ggplot2 conventions) and extended to the other stat_*() layers.label.strata parameter of stat_stratum() is deprecated in favor of infer.label, which is extended to the other stat_*() layers and sets label to alluvium in those cases rather than to stratum.aggregate.y parameter of stat_alluvium() is deprecated in favor of cement.alluvia, and the underlying procedure is debugged.aes.bind parameter of stat_flow() and stat_alluvium() now prefers character string options to logical values, described in the lode ordering vignette: "none", "flows", and "alluvia". The default "none" produces different behavior than the previous default FALSE, in that under this setting the aesthetic variables are not at all prioritized.weight and aggregate.wts are discontinued.Negative values can now be meaningfully passed to y, producing behavior that mimics that of geom_bar(). The new logical parameter absolute controls whether negative strata, and lodes within them, are ordered vertically in the same way as positive strata and lodes (FALSE) or in the opposite way (TRUE). Additionally, the negate.strata parameter can be used to negate the observations associated with specific strata, in order to situate them below rather than above the x axis.
The new lode guidance function lode_zagzig() mimics the behavior of lode_zigzag() except in initially “zagging” toward the farther end rather than “zigging” toward the closer end.
stat_*() internals have been simplified and standardized, in particular the manner in which lodes are ordered within strata.
Tests have been added for the statistical transformations. Visual regression tests using vdiffr have been added for the geoms.
The lode guidance functions have been renamed as follows and their original names retained as aliases:
| original | renamed |
|---|---|
| rightward | forward |
| leftward | backward |
| rightleft | frontback |
| leftright | backfront |
Additionally, lode.ordering now accepts a single integer vector of length the number of cases (alluvia), and will use the vector to sort the lodes within strata at each axis.
Finally, a new vignette showcases this and related functionality using a small example.
The following parameters, deprecated in previous versions, are now defunct (with informative messages):
weight in the stat_*()s (replaced with y)aggregate.wts in stat_alluvium() (replaced with aggregate.y)logical in the is_*_form()sThe width and knot.pos parameters sometimes required by Geom*$setup_data() are now set to the same defaults as in the geom_*()s when called from a stat. Previously-implemented warnings have been removed.
The lode.guidance argument of stat_alluvial() now accepts functions as input, making the use of custom functions easier as demonstrated in an example.
Parameters min.height and max.height are introduced to stat_stratum() to allow users to omit strata outside a given height range. This is probably most relevant for stratum labeling, as illustrated in the updated vignette.
session_info()Because the only functional (e.g. out README.md) occurrence of devtools is to call session_info() at the ends of the vignettes, this suggestion and usage are switched to sessioninfo.
Documentation is slightly reformatted due to switching roxygen syntax to markdown.
The internal z-ordering function z_order_aes failed to recognize contiguous segments of alluvia, thereby assigning later segments missing values of 'group' and preventing them from being rendered. This has been corrected.
geom_alluvium() patchAn occurrence of weight in geom_alluvium() was not updated for v0.8.0 and caused geom_alluvium() to throw an error in some cases. This has been corrected.
geom_flow() patchAn earlier solution to the z-ordering problem sufficed for matched layers (*_alluvium() and *_flow()) but failed for the combination of stat_alluvium() with geom_flow(). This is been corrected in the code for GeomFlow$draw_panel(), though a more elegant and general solution is preferred.
The deprecated parameters axis_width (all geom layers) and ribbon_bend (geom_alluvium() and geom_flow()) are removed and an explanatory note added to the layers’ documentation.
A vignette illustrating two methods for labeling small strata, using other ggplot2 extensions, is included.
self_adjoin() exportThe internal function self_adjoin(), invoked by geom_flow(), is revised, exported, documented, and exemplified.
weight aesthetic for the three stat_*() functions is replaced by the y aesthetic, so that scale_y_continuous() will correctly transform the vertical scales of the layers. An example is provided in the documentation for stat_alluvium(). The y aesthetic must be present in order for scales to be correctly transformed. The weight parameter is still available but deprecated.weight to y, the aggregate.wts parameter to stat_alluvium() is replaced with aggregate.y; aggregate.wts is deprecated.stat_*() functions.These changes make the functions that test for and convert between alluvial formats behave more like popular functions in the tidyverse. Some of the changes introduce backward incompatibilities, but most result in deprecation warnings.
is_alluvial_*() and to_*() are renamed to is_*_form() and to_*_form() for consistency. Their old names are deprecated.is_alluvial() is deprecated and will be removed in a future version.logical is deprecated. In a future version, the functions is_*_form() will only return logical values.silent = TRUE now silences all messages.is_*_form() now return FALSE if any weights are negative, with a message to this effect.key, value, id, weight, and diffuse parameters, using up-to-date rlang and tidyselect functionality.axes parameter in is_alluvia_form() and to_lodes_form() now accepts dplyr::vars() objects, as in dplyr::select_at(). Alternatively, variables can be fed to these functions as in dplyr::select(), to be collected by rlang::quos(...) and used as axis variables. If axes is not NULL, then such additional arguments are ignored.to_*_form() now merge their internal reshapen data frames with the distilled or diffused variables in a consistent order, placing the distilled or diffused variables to the left.Depends on R v3.3.0 (patch number zero) instead of v3.3.1. I’ve been unable to install this version locally, so there is a slight chance of incompatibility that i’ll be watchful for going forward.Suggests rather than Imports.is_alluvial_*() and to_*() functions are combined; see help("alluvial-data").is_alluvial_alluvia now prints a message rather than a warning when some combinations of strata are not linked by any alluvia.to_lodes() now has a diffuse parameter to join any original variables to the reformatted data by the id variable (alluvium). This makes it possible to assign original variables to aesthetics after reformatting, as illustrated in a new example.to_alluvia() now has a distill parameter to control the inclusion of any original variables that vary within values of id into the reformatted data, based on a distilling function that returns a single value from a vector.to_lodes() now has a logical discern parameter that uses make.unique() to make stratum values that appear at different axes distinct. The stat_*() functions can pass the same parameter internally and print a warning if the data is already in lodes form.GeomFlow$draw_panel() now begins by restricting to complete.cases(), corresponding to flows with both starting and terminating axes. (This is not done in StatFlow$compute_panel(), which would have the effect of excluding missing aesthetic values from legends.)GeomAlluvium$setup_data() now throws a warning if some color or differentiation aesthetics vary within alluvia.lode.ordering argument by StatAlluvium$compute_panel() has been fixed.The ggalluvial() shortcut function, which included a formula interface, deprecated in version 0.4.0, is removed.
I only started maintaining NEWS.md with version 0.5.0.