Open ElDeveloper opened 9 years ago
Do you have a specific example of two plots where you found the colors for the same taxon to be different? I had assumed that the coloring had remained consistent between visualizations, but then again I hadn't specifically investigated this across plots. Note that we're only recycling about ~16 colors in each visualization - we found that using more than that made the color gradients too hard to interpret.
I agree with limiting the number of colors.
The screenshots below show how Firmicutes has two different colors:
Thanks for those - are these two screenshots are from two different BIOM files that you independently uploaded? Right now the colors are re-assigned for each BIOM file you uploaded (based on the taxonomy/abundance information per file) - so if your rank abundance info for OTUs varied slightly between files, then the color assignments would differ. Definitely agree that we should fix this to avoid the need for separate keys.
Yes these are two independently uploaded tables. I agree, either that or being able to order the samples by different categories, for example by subject and then by timepoint so you end up with something like:
SA-1 SA-2 SA-3 SA-4 | SB-1 SB-2 SB-5 | SC-1 SC-10
Where you have two categories in the metadata HOST_SUBJECT_ID and Time-point.
On (Jan-28-15| 1:15), hollybik wrote:
Thanks for those - are these two screenshots are from two different BIOM files that you independently uploaded? Right now the colors are re-assigned for each BIOM file you uploaded (based on the taxonomy/abundance information per file) - so if your rank abundance info for OTUs varied slightly between files, then the color assignments would differ. Definitely agree that we should fix this to avoid the need for separate keys.
Reply to this email directly or view it on GitHub: https://github.com/PitchInteractiveInc/Phinch/issues/41#issuecomment-71802709
Currently there's not a good way to make two different plots comparable, because for each plot you make you get a different coloring scheme (i. e. each taxa is associated with a different color) and thus when putting together two or more of them you need to either manually fix the colors or provide two groups of color-taxa legends.