Aeolus: A Multi-structural Flight Delay Dataset

TL;DR

Aeolus addresses the gap between academic tabular benchmarks and real-world flight delay dynamics by introducing a unified, multimodal benchmark that combines tabular features, temporal flight chains, and graph-based flight networks. The dataset supports regression, classification, and uncertainty tasks with leakage-free temporal splits and provides baseline experiments across tabular, sequential, and graph modalities. Key contributions include three aligned data modalities, robust evaluation protocols, and methodological insights on distribution shifts and exogenous shocks such as COVID-19. The work enables more realistic, generalizable modeling for flight delays and offers a testbed for broader structured-data research in industrial settings.

Abstract

We introduce Aeolus, a large-scale Multi-modal Flight Delay Dataset designed to advance research on flight delay prediction and support the development of foundation models for tabular data. Existing datasets in this domain are typically limited to flat tabular structures and fail to capture the spatiotemporal dynamics inherent in delay propagation. Aeolus addresses this limitation by providing three aligned modalities: (i) a tabular dataset with rich operational, meteorological, and airportlevel features for over 50 million flights; (ii) a flight chain module that models delay propagation along sequential flight legs, capturing upstream and downstream dependencies; and (iii) a flight network graph that encodes shared aircraft, crew, and airport resource connections, enabling cross-flight relational reasoning. The dataset is carefully constructed with temporal splits, comprehensive features, and strict leakage prevention to support realistic and reproducible machine learning evaluation. Aeolus supports a broad range of tasks, including regression, classification, temporal structure modeling, and graph learning, serving as a unified benchmark across tabular, sequential, and graph modalities. We release baseline experiments and preprocessing tools to facilitate adoption. Aeolus fills a key gap for both domain-specific modeling and general-purpose structured data research.Our source code and data can be accessed at https://github.com/Flnny/Delay-data

Figures (4)

• Figure 1: Overview of Aeolus.
• Figure 2: Structure of Flight Chain.
• Figure 3: Structure of Flight Trees Constructed from Different Root Flights.
• Figure 4: Histogram and normal fit of ARR_DELAY and DEP_DELAY in 2024.