Airlines: Subsampled Flight Data for ML Benchmarking

A 2000-row, 100-column subsample of an airline operations dataset, generated with a specific random seed and class stratification. The dataset was created by Eddie Bergman using a Python script to uniformly sample rows, columns, and up to 10 target classes. Its original source is the 'airlines' dataset (ID 1169) on OpenML.

Use Cases

• Benchmarking classification model performance based on the 10 target classes.
• Feature selection experiments based on the 100 sampled columns.
• Testing the impact of stratified sampling on model evaluation.

Strengths

• Controlled subsampling with a specified random seed (2) ensures reproducibility.
• Stratified sampling was applied to maintain class distribution in the 2000-row subset.
• The dataset includes up to 100 features, providing a moderate feature space for analysis.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown for the original source dataset, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

OpenML dataset 'airlines' (ID 1169)

Collection Method

Algorithmic subsampling via a custom Python function to select rows, columns, and classes.