Airlines Departure Delay Prediction: 1 Million Flight Records for Regression Benchmarking

Bureau of Transportation Statistics' Airline Service Quality Performance data from 2013 provides a processed subset for predicting departure delays in seconds. This dataset is the first 1 million rows from a larger CSV covering years 1987 to 2013, transformed for a tabular regression benchmark. It is designed for regression tasks on both numerical and categorical features.

Use Cases

• Benchmarking regression model performance based on the dataset's stated role in a tabular data benchmark.
• Predicting flight departure delays in seconds based on the processed airline service data.
• Analyzing factors influencing airline on-time performance based on the original transportation statistics source.

Strengths

• Contains 1 million rows, providing a substantial sample for model training.
• Explicitly designed for a regression benchmark on mixed numerical and categorical features.
• Derived from an authoritative source, the U.S. Bureau of Transportation Statistics.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but the specific subset of columns used is unknown.
• Last update date is unknown; freshness unverified.

Provenance

Source

Bureau of Transportation Statistics, Airline Service Quality Performance

Collection Method

Processed version of original airline on-time performance data, subsetted and transformed for a benchmark.

Time Range

Original data source is from 2013; the underlying raw CSV covers 1987-2013.

Freshness

Source data is from 2013; the benchmark transformation date is unknown.

Geography

Likely United States, given the source is a U.S. government bureau, but not explicitly stated.