lizzieinvancouver / mastjc

0 stars 0 forks source link

make two conceptual figures #1

Open lizzieinvancouver opened 4 months ago

lizzieinvancouver commented 4 months ago

@ngoj1 Okay, I could use help with two separate figures sketched out here (and uploaded as PDF to figures folder): figureideas_13May2024

The basic point we're trying to show is that Janzen Connell (JC) and masting both make predictions for how seed density should affect pathogen/seed-predator pressure and thus seedling survival, but they differ totally in their predictions and what axes they focus on. JC is focused on space (specifically distance from the mother tree -- check out the wikipedia and see https://www.researchgate.net/figure/The-Janzen-Connell-hypothesis-of-seed-and-seedling-herbivory-PRC-population-recruitment_fig3_238668527 for a version of the usual figure people show to explain this) while masting is time (years since last mast seed year).

The figure figure (top) is just showing that the survival x seed density predictions vary. These are both vaguely non-linear with no exact underlying model, so we should just draw vaguely non-linear lines (also, it's not clear in masting what happens at low seed density times so perhaps dash half the line?) but it would be great if you can draw small simple pictures that capture each and we use those little images in the next figure. You can probably come up with good ideas for this using these, but I strongly suggest including some circle/cycle for masting (to represent time, I tried drawing 'mast year' then 'year + 1' and the 'year n' but not sure that works or there is space) and an arrow for JC (since time is well represented by an arrow).

The next figure is supposed to show three dimensions in 2D. I think of it as some sort of heat map, but see what you come up with. Basically, it tries to combine the time axis of masting with the space axis of JC -- there are TWO corners of the figure we know and the rest are sort of question marks. It would also be good to highlight here that each only looks at once axis, but I am not sure that will be easy (but that was my red and blue pen). See what you can come up with! I will aim to answer any queries but feel free to ping me on email if i am not responding here quickly.

Thank you!

ngoj1 commented 4 months ago

Hi Lizzie,

Here's a rough draft of the two figures: 2DFigure.pdf 3DFigure.pdf

The colours are just placeholders, but I think the viridis-esque colour looks pretty nice and I may use that for the 2D figure if you prefer it as well. For the 3D figure, is the heatmap meant to represent the pathogen/predator pressure? That's what I assumed when looking at the sketch, and in my head it makes sense that high seedling density caused by low dispersal distance and strong masting would promote the strongest predation due to seed abundance, and then low pressure when the seeds are more scattered away + fewer of them produced. For the question mark regions I've left them white for now.

lizzieinvancouver commented 4 months ago

@ngoj1 These are FANTASTIC! I really like them. I have a few tweaks (and my co-authors may as well). I am about to be in a meeting though so will send those later... for now:

For the 3D figure, is the heatmap meant to represent the pathogen/predator pressure? That's what I assumed when looking at the sketch Yes!

I wonder if would make some of the squares on the heat map stippled or dotted (or hashed)? We could do that in lieu of question marks; if possible it would be nice. Or perhaps we can overlay small white dots or such....

As for the colors, I like them but also consider the other viridis palettes and pick whichever seems most like it says pathogen/predator pressure to you.

lizzieinvancouver commented 4 months ago

Okay, the small tweaks!

Thank you!

ngoj1 commented 4 months ago

Hi Lizzie,

Here's an update on the draft figures. 2D Figure: image

3D Figure, version 1: Dithering/pixellating image

3D Figure, version 2: Halftone/circles image

For both the 2D and 3D Figures, the colour scheme is based off viridis' "Plasma" as I thought that the red is quite bold and invokes some feeling of danger (?).

If you're happy with either option I am ok to leave them as is, but something seems a bit off about the stippling/white gradient to me personally...I can try another method that may look better but it takes a bit more time (maybe by the end of the weekend?), so if you have a preference between the pixels or the circles just let me know and I'll try to tweak them.

I also considered swapping the colour key around since right now High is at the bottom, where Low would usually be, but it just aligns with the colours on the heat map too well in the current orientation, but I can easily change that if you want.

These are all saved as .ai and .pdf files in the most recent push that I sent out, under figures/justinDraft.

ngoj1 commented 4 months ago

Actually I figured it out, it was the circle spacing. Here's what I personally think looks the best among the 3 versions of Figure 3D: image

Let me know what you think!

lizzieinvancouver commented 4 months ago

@ngoj1 These look amazing to me! I don't have any tweaks now, but need to think on where the stippling should be (definitely where we have, but want to confirm with coauthors about the middle of the figure). Super cool! Did you do the stippling in R?

ngoj1 commented 4 months ago

I took a public use halftone gradient from Google that was originally in black and just manipulated the image in Illustrator to make it transparent background + invert colour before overlaying it manually on the corners of the heatmap and then trimming away the excess. I wish I knew a more elegant solution that involved coding, but I'm just more familiar with this kind of freehand graphic design.

lizzieinvancouver commented 4 months ago

@ngoj1 Awesome! I suspect there is no easy way to do that with coding, especially so elegantly, so it's great you know how to do it otherwise.