examples/dataset_explorer.py

import marimo

__generated_with = "0.9.18"
app = marimo.App(width="full")


@app.cell
def __():
    datasets = [
        # Add your own HF datasets
        "scikit-learn/iris/Iris.csv",
        "scikit-learn/adult-census-income/adult.csv",
        "scikit-learn/auto-mpg/auto-mpg.csv",
        "scikit-learn/credit-card-clients/UCI_Credit_Card.csv",
        "scikit-learn/Fish/Fish.csv",
        "scikit-learn/tips/tips.csv",
    ]
    return (datasets,)


@app.cell(hide_code=True)
def __(mo):
    mo.md(r"""## Select a dataset""")
    return


@app.cell(hide_code=True)
def __(datasets, mo):
    dataset = mo.ui.dropdown(datasets, value=datasets[0], label="Select a dataset")
    no_limit = mo.ui.switch(label="Limit 1000", value=True)
    mo.hstack([dataset, no_limit])
    return dataset, no_limit


@app.cell
def __(dataset, mo, no_limit):
    explore = mo.sql(
        f"""
        CREATE OR REPLACE TEMP TABLE explore
        AS (FROM 'hf://datasets/{dataset.value}')
        {'LIMIT 1000' if no_limit.value else ''};

        FROM explore;
        """
    )
    return (explore,)


@app.cell(hide_code=True)
def __(mo):
    mo.md(r"""## Summary""")
    return


@app.cell(hide_code=True)
def __(explore, mo):
    _schema = mo.accordion({"Schema": explore.schema})

    mo.md(f"""
    * Total rows: **{len(explore):,}**
    * Total columns: **{len(explore.columns)}**

    {_schema}
    """)
    return


@app.cell
def __(explore):
    explore.describe()
    return


@app.cell(hide_code=True)
def __(mo):
    mo.md("""## Manipulate the data""")
    return


@app.cell
def __(explore, mo):
    transformed = mo.ui.dataframe(explore)
    transformed
    return (transformed,)


@app.cell(hide_code=True)
def __(mo):
    mo.md(r"""## Explore the data""")
    return


@app.cell
def __(mo, transformed):
    mo.ui.data_explorer(transformed.value)
    return


@app.cell(hide_code=True)
def __():
    # Imports
    import marimo as mo
    import polars
    return mo, polars


if __name__ == "__main__":
    app.run()