Allotaxonograph demo

Compute the rank-turbulence divergence between the bundled climate and finance corpora, list the top contributors, and save the allotaxonograph.

What it shows

1. Building two RankedLists from the demo corpus pair
2. The rank-turbulence divergence and its top per-type contributions
3. Saving the two-panel allotaxonograph to a PNG

Run it

python examples/allotaxonograph_demo.py

Rank-turbulence divergence (alpha=1/3): 0.267
Top contributors:
  climate      0.036  (climate)
  market       0.036  (finance)
  carbon       0.030  (climate)
  stock        0.030  (finance)
  trade        0.026  (finance)

Saved allotaxonograph.png

Source

examples/allotaxonograph_demo.py

"""Allotaxonograph demo: rank-turbulence divergence and the rank-rank map.

Run with:  uv run python examples/allotaxonograph_demo.py
Writes allotaxonograph.png to the current directory.
"""

from keyflux import RankedList, allotaxonograph, rtd
from keyflux.datasets import load_demo_pair


def main() -> None:
    """Compare the bundled climate and finance corpora and save the figure."""
    focus, reference = load_demo_pair()
    climate = RankedList.from_counts(focus, label="climate")
    finance = RankedList.from_counts(reference, label="finance")

    result = rtd(climate, finance, alpha=1.0 / 3.0)
    print(f"Rank-turbulence divergence (alpha=1/3): {result.divergence:.3f}")
    print("Top contributors:")
    for contribution in result.contributions[:5]:
        leans = "climate" if contribution.direction == "system1" else "finance"
        print(f"  {contribution.type:<12} {contribution.contribution:.3f}  ({leans})")

    fig = allotaxonograph(climate, finance, alpha=1.0 / 3.0)
    fig.savefig("allotaxonograph.png", dpi=120)
    print("\nSaved allotaxonograph.png")


if __name__ == "__main__":
    main()