plot_weibull()

Plot Kaplan-Meier estimates on distribution-linearizing axes.

Usage

Source

plot_weibull(
    km,
    *,
    aft=None,
    dist="weibull",
    title=None,
    xlab=None,
    ylab=None,
    width=500,
    height=300,
)

Produces a diagnostic scatter plot of Kaplan-Meier survival estimates on transformed axes where data from the specified parametric distribution falls on a straight line. This is the classic “Weibull plot” (complementary log-log transform) generalized to other AFT distributions.

For stratified Kaplan-Meier fits, each stratum is drawn as a separate colored series. Parallel lines indicate the proportional hazards assumption holds between groups.

An optional fitted AFT model overlays the theoretical parametric line on the empirical points, showing goodness-of-fit visually. The overlay is available only for unstratified Kaplan-Meier fits.

Parameters

km: KaplanMeier

A fitted KaplanMeier object (stratified or unstratified).

aft: Any | None = None

An optional fitted AFT model for a parametric overlay line. Only allowed with an unstratified km. The AFT’s distribution must match the dist parameter (with "exponential" accepted when dist="weibull").

dist: str = "weibull"

Distribution whose linearizing transform is applied: "weibull" (default), "lognormal", or "loglogistic".

title: str | None = None

Chart title. Defaults to None (no title).

xlab: str | None = None

X-axis label. Defaults to "log(time)".

ylab: str | None = None

Y-axis label. Defaults to a distribution-specific label.

width: int = 500

Chart dimensions in pixels (defaults 500 x 300).

height: int = 500

Chart dimensions in pixels (defaults 500 x 300).

Returns

altair.LayerChart

An interactive Altair chart.

Examples

Check the Weibull assumption for lung cancer survival by sex:

import greenwood as gw

lung = gw.load_dataset("lung", backend="polars")
death = gw.Outcome.surv(time="time", event="status")

# Stratified KM where parallel lines suggest PH holds
km = gw.KaplanMeier().fit(death, by="sex", data=lung)
gw.plot_weibull(km)

Overlay a fitted Weibull AFT on an unstratified KM:

km_overall = gw.KaplanMeier().fit(death, data=lung)
aft = gw.AFT(dist="weibull").fit("Surv(time, status) ~ age + sex", data=lung)
gw.plot_weibull(km_overall, aft=aft)