pyrolite.plot.density.ternary

pyrolite.plot.density.ternary.ternary_heatmap(data, bins=20, mode='density', transform=<function ILR>, inverse_transform=<function inverse_ILR>, ternary_min_value=0.0001, grid_border_frac=0.1, grid=None, **kwargs)[source]

Heatmap for ternary diagrams. This invokes a 3D to 2D transform such as a log transform prior to creating a grid.

Parameters

  • data (numpy.ndarray) – Ternary data to obtain heatmap coords from.

  • bins (int) – Number of bins for the grid.

  • mode (str, {'histogram', 'density'}) – Which mode to render the histogram/KDE in.

  • transform (callable | sklearn.base.TransformerMixin) – Callable function or Transformer class.

  • inverse_transform (callable) – Inverse function for transform, necessary if transformer class not specified.

  • ternary_min_value (float) – Optional specification of minimum values within a ternary diagram to draw the transformed grid.

  • grid_border_frac (float) – Size of border around the grid, expressed as a fraction of the total grid range.

  • grid (numpy.ndarray) – Grid coordinates to sample at, if already calculated. For the density mode, this is a (nsamples, 2) array. For histograms, this is a two-member list of bin edges.

Returns

  • t, l, r (tuple of numpy.ndarray) – Ternary coordinates for the heatmap.

  • H (numpy.ndarray) – Histogram/density estimates for the coordinates.

  • data (dict) – Data dictonary with grid arrays and relevant information.

Notes

Zeros will not render in this heatmap, consider replacing zeros with small values or imputing them if they must be incorporated.