The outliers were being incorrectly excluded when computing the quartiles. I've
also added a +/-1.5 IQR whiskers computation for the Morley-Michelson example,
so it replicates the R plot exactly.
This specifies a function that takes the sorted data array, and returns an array
of datum positions that should marked with whiskers. The default implementation
is to return `[0, length-1]` i.e. the minimum and maximum.
Data outside of the whiskers are considered outliers, and are not included in
the quartile calculation.
The pack layout now sorts nodes by ascending value by default. The `radius`
property is removed, as the radius is always computed from the node value. The
`spacing` property is also removed; it doesn't seem particularly useful. The
`radius` property is renamed to `r` to match `x` and `y`.
I also optimized the implementation slightly, and namespaced temporary state
under the prefix `_pack_`.
Based on the Protovis version. The only difference is that I've dropped
the `orient`, `innerRadius` and `outerRadius` properties so that the D3
version is more flexible.
Based on the Protovis layout. I've kept the convention of using `size`
for the width/height and `separation` for the spacing. The Protovis
layout had a `size` property but this was used to compute the radius.
I've left this out as I've assumed it's straightforward enough to pass
in the appropriate `radius` function instead. Likewise, I think the
existing hierarchy `sort` property can handle situations that `order`
was intended for in the Protovis version.
Rather than specifying an orientation, compute the layout in normalized space. In a future change,
I'd also like to clean up the temporary state that is stored on tree nodes, and record dx and dy
attributes per node that could be used to compute a space-filling layout.
We were computing the tick join based on the (new) format function, but doesn't
produce the desired effect: the new format is applied on the old data. Thus, the
wrong join occurs if, say, the value 0.5 with the new format results in "0". The
correct join compares the old text content to the new format value.
This example shows how to preserve scales across multiples using a single chart
instance, as opposed to bullet-multiples which uses distinct chart instances to
supply separate scales. I still think it'd be better to use a single chart
instance in both cases, but then I need a different place to hide the scale
state.
This way, we get separate scales for the small multiples, which makes sense
given our data. However, I'm not totally convinced this is the right way to
implement separate scales, because it's a bit awkward to create separate chart
instances that look identical. Also, it's unfortunate that the charts are
stateful; it'd be better to somehow store the scale as data on the nodes, so
that chart specifications could be more easily reused. Then, there might be a
method to fix the domain rather than computing the domain per-chart.
It's nice, but I think it's a bit more flexible to not have it as part of the
chart specification. This way, people can define titles however they like. It
might be nice to take a similar approach with reference ticks in the future.
We now preserve object constancy for ticks across transitions. By caching a
reference to the previous x-scale, we can initialize entering objects in the
correct location, then transition them to the new scale as they fade in. Also,
we use the `map` operator to convert the data to a standard representation that
is suitable for the bullet chart, and compute derivate data needed across
multiples.
The treemap layout can now be set to "sticky" mode, which preserves the row
arrangements from an earlier layout. This is particularly handy for animating
treemaps, as it avoids the temporary occlusion as nodes swap positions. Thanks
to @philogb and the New York Times for the inspiration!
Still a work in progress. Need to decide on the appropriate API for representing
tree structures; not sure I want to go the heavyweight pv.dom route. May also
need rounding to exact pixel values.
Use the Floyd-Warshall algorithm to compute the shortest path between nodes, and
use that graph theoretic distance as the distance constraint for the Gauss-
Seidel relaxation. In addition, have the constraint alpha decay over time, as in
Simulated Annealing.
This fixes a bug in the pie layout sorting where we want the sort order to
affect the layout, but not the order in which arcs are rendered—and furthermore
we want the order of arcs to always match the order of data. (If you want to
sort the data, do that before it is passed to the layout.)
This adds d3.layout.pie, which can be used to compute the start and end angles
for arcs given an arbitrary array of data. In addition, d3.svg.arc now has a
`centroid` method that computes the center of the arc, useful for labeling.
Together these changes greatly simplify the donut & pie examples.
This commit also includes a `map` method for selections, which is similar to the
`filter` method. This allows you to map the data bound to the current selection.
I'm not currently using it, but it seems like a useful feature for those cases
where you want to rebind the current selection to different (but related) data.
D3 now supports Sizzle, preferring Sizzle to the native Selectors API if Sizzle
is available. Sizzle internally uses the native Selectors API and thus this
should have minimal performance implications; also, it allows you to use Sizzle
proprietary extensions such as ":first".
This commit also restricts the definition of the enter selection so that only
append and insert operations are defined. The other operations were generally
unsupported anyway, and it cleans up the code to have separate implementations
for insert and append. (I might enable additional operations in the future, such
as `filter`, `sort` and `each`, but this seems like a reasonable first pass.)
The `symbol` class can be used to generate dot symbols for scatterplots. This
commit also fixes a couple bugs with negative-domain log and pow scales, as well
as tick generation.
This method can be used for computing the bounding box of arbitrary GeoJSON
objects. This commit also fixes a bug in d3.geo.path, such that it will accept
any GeoJSON object, rather than requiring GeoJSON feature objects.
This uses d3.geo.path to compute the projected area (in square pixels) of each
county in the choropleth. It also usees d3.behavior.zoom to add geometric pan
and zoom.
Includes, as the first behavior, a pan & zoom behavior. The canvas can be panned
by dragging the mouse, and zoomed using the mousewheel (or by double-click). By
listening to redraw events, users can decide whether to implement geometric
zooming (such as by setting the "transform" attribute on an `svg:g` element) or
semantic zooming (by changing the domain of a scale object and repositioning
elements).
This commit also includes two bug fixes. The `d3.format` class now properly
groups thousands of negative numbers, and supports the sign specifier. The
unicode minus symbol \u2212 is used for negative values. The `d3.scale.pow`
class now properly handles negative numbers, as well.
First, always coerce the tween target value (the value of the style or attribute
at the end of the transition) to a string. This guarantees that the correct
interpolator is chosen, and avoids a null dereference.
Second, default color channels to zero rather than undefined, so as to avoid NaN
channel values if there is an error parsing the color. This can occur, say, if
the background color of a div is unset and defaults to "rgba(0, 0, 0, 0)".
The data join is now specified as a single function of data, as with all other
properties. This allows the key to be computed on the previously-bound data,
rather than requiring the key to be serialized into the DOM (say, as an
attribute). In the case that there is no previously-bound data, it is still
possible to access the associated node as the `this` context.
The `enter` operator no longer performs an append. For symmetry with the `exit`
operator, you must call `append` after obtaining the entering selection. This
requires a tiny bit more code, but should make the code more clear. Also, it
provides an opportunity to use a different instantiation operator, such as the
new `insert` operator. This takes a second argument, which is a selector for the
insert-before reference element. For example, the selector ":first-child" will
prepend nodes.
The `empty` operator allows you to query whether a selection is empty (i.e.,
contains zero matching nodes).
Jason Davies
Mike Bostock