Year-in-Review Sunburst Chart does not incorporate observations of non-leaf taxa

[pisum]

The end of the year is approaching. So i thought i’d officially log a small bug in the sunburst chart in the year-in-review page that I had noted earlier in the year:

@pisum

it’s not counting observations that are at taxa higher than species in cases where those taxa are not the “leaves”. so the counts of taxa that are harder to identify to species (like fungi and flies) seem to be undercounted in general

For example, here is the sunburst chart from tiwane’s 2020 YIR page, zoomed in on spiders:

Note that when I highlight over the block that represents Lycosoidea, it the chart says there are 3 observations, and if you follow that out, you would see that that represents 1 observation each of genera Pardosa, Oxyopes, and Hamadruas.

But if you search for tiwane’s actual observations of Lycosoidea, you’ll find 3 additional observations at family and higher, for a total of 6 Lycosoidea observations:

Ideally, that YIR sunburst chart would account for these observations at non-leaf taxa. In this case, it should show a block of 2 observations for Lycosidae observation at family level (displayed at the same level as the genus blocks) and another block with a single observation that represents the Lycosoidea observation at superfamily level (displayed at the same level as the family level blocks). And then all the parent block counts would be bumped up accordingly.

[tiwane]

We released a fix - stats will need to be regenerated for it to be applied, however.

[pisum]

i don’t think the new code works like i would expect. it is now displaying the correct numbers when you highlight over each block, but the size of the blocks don’t accurately reflect the numbers. below, i’ve zoomed into your Lycosoidea block:

i added numbers that represent the counts that are displayed when you hover over each block. the numbers are correct. however, the size of the blocks aren’t proportionate to the counts. for example, note that the Wolf Spiders (count=3) block is half the size of Lynx Spiders (count=2) block. Wolf spiders has 1.5 times the count vs Lynx Spiders. so it should be 1.5 times as large, not 0.5 times.

here’s roughly what i would expect to see instead:

or (if you display a block representing observations at the parent rank):

@pisum

it should show a block of 2 observations for Lycosidae observation at family level (displayed at the same level as the genus blocks) and another block with a single observation that represents the Lycosoidea observation at superfamily level (displayed at the same level as the family level blocks). And then all the parent block counts would be bumped up accordingly.