get_missing_summary() helper method for OpenML datasets

Start at the OpenMLDataset class and inspect its existing get_data() entry point. Implement the requested missing-value summary using the dataset's returned dataframe, including total and per-column counts, then verify that the result matches the example structure and values for a dataset with missing data.

Hi OpenML team,

I’d like to propose a small usability enhancement to improve the dataset exploration workflow in openml-python.

Feature Request

helper method:

dataset.get_missing_summary()

• that returns a simple summary of missing values for the dataset.

Motivation

Many users load a dataset and immediately need to check:

• how many missing values exist in total
• which columns contain missing values
• basic column-level counts

Currently, users must manually compute this after calling dataset.get_data().
A built-in helper would reduce repetitive code and improve the dataset exploration experience, especially for new users.

Proposed Behavior

dataset = openml.datasets.get_dataset(dataset_id)
df, *_ = dataset.get_data()

dataset.get_missing_summary()

Example output:

{
    "n_missing_total": 235,
    "missing_per_column": {
        "age": 10,
        "income": 20,
        "zipcode": 205
    }
}

Implementation Idea

• Implement this as a method inside the OpenMLDataset class.
• Internally, the method would:

1. Call .get_data()
2. Compute missing summary using pandas (df.isna().sum())
3. Return a dictionary with overall and per-column counts

No changes needed to the core API; this is an isolated helper.

Benefits

Improves ease of use

• No backward compatibility impact
• Lightweight feature (easy to maintain)
• Helps users performing initial dataset checks

Happy to open a PR implementing this.