Plotting models

The SDMs are integrated with the Makie plotting package.

using SpeciesDistributionToolkit
using CairoMakie

We will use the demonstration data from the SDeMo package:

model = SDM(RawData, NaiveBayes, SDeMo.__demodata()...)

❎  RawData → NaiveBayes → P(x) ≥ 0.5 🗺️

Plotting instances

f = Figure()
ax = Axis(f[1, 1]; aspect = DataAspect())
scatter!(ax, model)
current_figure()

This can be coupled with information about the model itself, to provide more interesting visualisations:

f = Figure()
ax = Axis(f[1, 1]; aspect = DataAspect())
scatter!(ax, model; color = labels(model))
current_figure()

Note that models also implement the occurences interface, so we can easily split presence and absences from the model.

f = Figure()
ax = Axis(f[1, 1]; aspect = DataAspect())
scatter!(
    ax,
    presences(model);
    color = :white,
    strokecolor = :orange,
    strokewidth = 1,
    label = "Presences",
)
scatter!(
    ax,
    absences(model);
    color = :white,
    strokecolor = :teal,
    strokewidth = 0.5,
    markersize = 5,
    label = "Absences",
)
axislegend(ax; position = :lt)
hidespines!(ax)
hidedecorations!(ax)
current_figure()

The models currently do not have a field for the species name, but this will be part of a future release. If you need to specify the species name from a model, for now you need to use Occurrences(model; name="Sitta whiteheadi").

Model diagnostic plots

These plots require a trained model:

train!(model)

☑️  RawData → NaiveBayes → P(x) ≥ 0.034 🗺️

Coming soon

These plots will be included in a future release of SDeMo- stay tuned!