How to Implement Your Own Visualization Function

In this section, we show how to implement your own visualization function. Here, we implement a visualization function that plots the optimization history.

First of all, import optuna and other required modules.

from __future__ import annotations

import matplotlib
import matplotlib.pyplot as plt
import optuna
import plotly.graph_objects as go

If you use plotly for visualization, the function should return a plotly.graph_objects.Figure object.

def plot_optimizaiton_history(study: optuna.study.Study) -> go.Figure:
    trials = study.trials
    values = [trial.value for trial in trials]
    best_values = [min(values[: i + 1]) for i in range(len(values))]  # type: ignore
    iterations = list(range(len(trials)))

    fig = go.Figure()
    fig.add_trace(
        go.Scatter(
            x=iterations,
            y=best_values,
            mode="lines+markers",
        )
    )
    fig.update_layout(title="Optimization history")
    return fig

If you use matplotlib for visualization, the function should return a matplotlib.figure.Figure object.

def plot_optimizaiton_history_matplotlib(study: optuna.study.Study) -> matplotlib.figure.Figure:
    trials = study.trials
    values = [trial.value for trial in trials]
    best_values = [min(values[: i + 1]) for i in range(len(values))]  # type: ignore

    fig, ax = plt.subplots()
    ax.set_title("Optimization history")
    ax.plot(best_values, marker="o")
    return fig

Plot the optimization history using the implemented visualization function. Here, we use the simple quadratic function as objective function.

def objective(trial: optuna.trial.Trial) -> float:
    x = trial.suggest_float("x", -10, 10)
    return x**2


study = optuna.create_study()
study.optimize(objective, n_trials=100)

fig = plot_optimizaiton_history(study)
fig.show()  # plt.show() for matplotlib

After implementing your own visualization function, you can register it with OptunaHub. See How to Register Your Algorithm with OptunaHub for how to register your visualization function with OptunaHub.