Website by @wy_bio to explore iNaturalist data - feedback wanted

General
82

I haven’t been on the forum for a little while and missed this, but it looks great and is a highly compelling use of phenology outputs! Working with Keir’s idea, I looked at Valley oak hybrids briefly to overlay and see how different flowering period of Valley oak compares to species it is known to hybridize with.

image
image1862×917 121 KB

The graph is really interesting, if the preliminary results are correct, blue oak generally peaks in flowering after valley oak (with a broader peak overall), and their hybrid (Jolon oak) has an even wider flowering period than either parent, though it may primarily peak right when valley oak does!

I think the only suggestions I have are relatively minor and maybe tricky to implement, but:

  1. custom/selectable colors (or at least consistent color choice defaults–e.g. 2nd output is blue) would be very helpful for comparing between graph outputs. For example, when comparing the graph output above with that for Q. garryana garryana x lobata, it’d be nice to have the colors match for valley oak and hybrid species.

In the graph below, it’s interesting to note that the flowering period for Q. garryana garryana x lobata (as for Q. douglasii x lobata) is broadened, but the peak is reversed here, being later than both parents! This may be related to a less broad distribution generally, but I’d have to hypothesis test by including only selected counties as Keir also did above.

image
image1868×931 127 KB

and related to above (if not implementable)

  1. movable taxa order and consistent color selection (e.g. 1st species is blue, 2nd red, 3rd purple–mix of blue and red!)

It’d be conceptually easier to work with, and custom colors would allow choice to have consistent colors between parents and hybrid, or to have consistent color between one parent and custom choice for the other parent and hybrid if that all makes sense.

Again, super cool and very happy with the displays already, but would also really appreciate a couple tweaks mentioned above.

@truthseqr would probably also be interested in the outputs above and visualization capabilities generally for these since she’s put so much great work into annotating thousands of oak observations!

83

Here’s a 16 color color-blind palette that you can try. I ran out of colors just comparing some related species on the map, so it really would be great to have more colors.

84

Line thickness could work on the graphs, but it doesn’t work on the maps. I need a consistent way to represent species that works for both maps and graphs,

Color blindness is just one of the issues need to be dealt with when choosing colors for data visualizations. Another issue is dealing with how people’s perception of color affects how they interpret the data. There are people who study color theory for data visualizations like Cynthia Brewer and Fabio Crameri. While the color-blind palette that you linked to might be good for graphic design, it might not be good for data visualization where color is used to communicate information.

85

However, it is better than using one color to compare two things, which is much worse for data visualization. It would be cool if the website could use a preferred palette when the number of taxa compared fits but switch to a larger and maybe less preferred palette when the preferred palette numbers are exceeded. From a data visualization standpoint, comparing less is clearer but sometimes you want to compare more.