Reads a boxplot as the five-number summary it hands over.
Measured, the series carries ['base', 'min', 'Q1', 'median', 'Q3', 'max']
with one row per box and getName(i) answering the category, so the
reading is a transcription -- nothing is derived from the drawing.
It is read without an outline, and that is a property of the drawing
rather than a gap in the measurement.BoxSelector wants one selector
for each part -- the whiskers, the box, the median, each outlier -- and ECharts
draws all of them as one path. Colour-tagged, a two-box chart yields
exactly two filled paths, and each one's d is the box rectangle, a Z,
and then the whiskers:
There is nothing to name the parts with. On top of that the default paint
is #fff, which this adapter's filter counts as furniture, so the mark
would be dropped even if the shape fitted. Reading without an outline is
what gauge, sankey, graph, parallel and themeRiver already do
here: the values, the text and the braille all still work; only the
highlight is absent.
Outliers are empty rather than guessed. ECharts draws them as a
separate scatter series by its own convention, so a boxplot series
carries none of them -- there is no seventh dimension holding them. An
accompanying scatter is read as the scatter it is.
Reads a
boxplotas the five-number summary it hands over.Measured, the series carries
['base', 'min', 'Q1', 'median', 'Q3', 'max']with one row per box andgetName(i)answering the category, so the reading is a transcription -- nothing is derived from the drawing.It is read without an outline, and that is a property of the drawing rather than a gap in the measurement.
BoxSelectorwants one selector for each part -- the whiskers, the box, the median, each outlier -- and ECharts draws all of them as one path. Colour-tagged, a two-box chart yields exactly two filled paths, and each one'sdis the box rectangle, aZ, and then the whiskers:There is nothing to name the parts with. On top of that the default paint is
#fff, which this adapter's filter counts as furniture, so the mark would be dropped even if the shape fitted. Reading without an outline is whatgauge,sankey,graph,parallelandthemeRiveralready do here: the values, the text and the braille all still work; only the highlight is absent.Outliers are empty rather than guessed. ECharts draws them as a separate
scatterseries by its own convention, so a boxplot series carries none of them -- there is no seventh dimension holding them. An accompanying scatter is read as the scatter it is.