Layered lollipop plot

Combine rules and point marks to show positive and negative values.

Rules form the stems and point marks form the heads. Upward and downward triangles distinguish positive and negative values; diamonds mark values near zero. A separate rule marks the zero baseline.

Data use and provenance

NumPy creates regularly spaced positions; GenomeSpy calculates the sine wave. No source dataset is required.

Adapted from the official GenomeSpy example.

Code

"""Layered lollipop plot.

Combine rules and point marks to show positive and negative values.
"""

import numpy as np
import pandas as pd

import genome_spy as gs


# Generate regularly spaced positions along a sine wave.
data = pd.DataFrame({"x": np.arange(0, 6.284, 0.39269908169)})

baseline = (
    gs.Chart(gs.Data(values=[{"zero": 0}]))
    .mark_rule(color="lightgray")
    .encode(y=gs.Y("zero:Q").title(None))
)
stems = gs.Chart().mark_rule(size=3)

# Point upward or downward according to the sign; use diamonds near zero.
heads = (
    gs.Chart()
    .mark_point(size=500, filled=True)
    .encode(
        shape=gs.Shape("sin(x):N")
        .scale(
            type="threshold",
            domain=[-0.01, 0.01],
            range=["triangle-down", "diamond", "triangle-up"],
        )
        .legend(None)
    )
)
lollipops = (
    (stems + heads)
    .properties(data=data)
    .transform_calculate(**{"sin(x)": gs.expr.sin(gs.datum.x)})
    .encode(
        x=gs.X("x:Q"), y=gs.Y("sin(x):Q").scale(padding=0.1), color=gs.Color("sin(x):Q")
    )
)
chart = baseline + lollipops

Show generated specification