search

BrunaLab / HeliconiaDemography

Effects of drought and habitat fragmentation on Heliconia acuminata.
Other
0 stars 1 forks source link

Plot raw data more when possible #79

Closed Aariq closed 3 years ago

Aariq commented 3 years ago

R1:

Related to above, I think it’s important to see more of the raw data than are presented in Fig. 1. (I’m thinking of a version of Fig 3 that shows size in t+1, survival, and flowering rather than effect size on the y-axis, with the data and lines showing the shape of the GAMs). This could be an appendix, but it would help with interpreting Figures 3 – 6.

I think this suggestion is to revert to plotting back-transformed fitted data and adding points for raw data. I think this is probably a good idea. Rather than just adding the model intercept though, it would probably be better to predict() for an median sized plant and use the same size as newdata for both CF and FF.

Related to #78

Aariq commented 3 years ago

Adding raw data too busy for main text, I think. Needs to be separate plots for CF and FF to be readable. Screen Shot 2021-08-23 at 12 55 24 PM

Violin plots for probabilities look OK Screen Shot 2021-08-23 at 12 42 15 PM Screen Shot 2021-08-23 at 12 41 56 PM

Aariq commented 3 years ago

Probably could do something cool with this: https://github.com/LKremer/ggpointdensity

Aariq commented 3 years ago

The best thing might be to plot CF and 1-ha in separate plots with raw data using ggpointdensity for size and using histograms or violin plots (as above) for flowering and survival. This should be in supplemental though, as the main point of these plots is to compare habitats on the same plot.