跳转至

文章背景与核心概要

本篇文章介绍了一项发表于 arXiv 的最新研究，推出了首个公开的大规模芬兰铁路列车运营与气象同步观测数据集——FI-TW（Finland Integrated Train-Weather）。该数据集涵盖了 2018 年至 2024 年期间的数据，通过哈弗辛距离（Haversine distance）进行时空对齐，整合了芬兰全境 5,915 公里铁路线路上的约 3,850 万条观测数据。

在技术核心方面，FI-TW 数据集融合了来自“芬兰数字化交通铁路服务”（Finland Digitraffic Railway Traffic Service）的运营指标与 209 个环境监测站的气象测量数据，共包含 28 个经工程处理的特征。研究团队通过空间回退算法、时间特征的循环编码以及气象数据的鲁棒缩放，有效解决了数据缺失和传感器异常值问题。基于该数据集的基线实验表明，利用 XGBoost 回归模型预测特定车站的晚点时间，平均绝对误差（MAE）仅为 2.73 分钟。

该研究的重大意义在于打破了以往气象数据与运营数据相互隔离的局限。它不仅揭示了显著的季节性和地理趋势（如冬季晚点率超 25%、中北部走廊高晚点聚集），还为机器学习在铁路运营研究中的应用提供了坚实支撑，广泛适用于列车晚点预测、天气影响评估以及铁路基础设施脆弱性测绘等领域。

FI-TW: An Open Train-Weather Dataset for Railway Delay Analysis in Finland

FI-TW: An Open Train-Weather Dataset for Railway Delay Analysis in Finland

Summary

Summary

The FI-TW (Finland Integrated Train-Weather) dataset is the first publicly available, large-scale resource combining Finnish railway operations with synchronized meteorological observations spanning from 2018 to 2024. Designed for machine learning and operational research, the dataset integrates around 38.5 million observations across Finland's 5,915-kilometer rail network using spatial-temporal alignment via Haversine distance. It includes 28 engineered features that bridge operational metrics (from the Finland Digitraffic Railway Traffic Service) with weather measurements (from 209 environmental monitoring stations).

The FI-TW (Finland Integrated Train-Weather) dataset is the first publicly available, large-scale resource combining Finnish railway operations with synchronized meteorological observations spanning from 2018 to 2024. Designed for machine learning and operational research, the dataset integrates around 38.5 million observations across Finland's 5,915-kilometer rail network using spatial-temporal alignment via Haversine distance. It includes 28 engineered features that bridge operational metrics (from the Finland Digitraffic Railway Traffic Service) with weather measurements (from 209 environmental monitoring stations).

Key findings and baseline experiments using the dataset reveal: - Seasonal and Geographic Trends: Winter months experience delay rates exceeding 25%, alongside distinct geographic clustering of high-delay corridors in central and northern Finland. - Machine Learning Utility: A baseline XGBoost regression model achieved a Mean Absolute Error (MAE) of 2.73 minutes for predicting station-specific delays. - Applications: The dataset supports train delay prediction, weather impact assessments, and railway infrastructure vulnerability mapping.

Key findings and baseline experiments using the dataset reveal: - Seasonal and Geographic Trends: Winter months experience delay rates exceeding 25%, alongside distinct geographic clustering of high-delay corridors in central and northern Finland. - Machine Learning Utility: A baseline XGBoost regression model achieved a Mean Absolute Error (MAE) of 2.73 minutes for predicting station-specific delays. - Applications: The dataset supports train delay prediction, weather impact assessments, and railway infrastructure vulnerability mapping.

Metadata

Metadata

Field Details
arXiv ID arXiv:2601.16592 (cs.LG)
Subjects Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Databases (cs.DB)
Authors Vinicius Pozzobon Borin, Jean Michel de Souza Sant'Ana, Usama Raheel, Nurul Huda Mahmood
Submitted / Revised Submitted on 23 Jan 2026; Last revised 6 Aug 2026 (v2)
Dataset Link Kaggle Dataset Repository
Full-Text Links View PDF | TeX Source | HTML Version
Field Details
arXiv ID arXiv:2601.16592 (cs.LG)
Subjects Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Databases (cs.DB)
Authors Vinicius Pozzobon Borin, Jean Michel de Souza Sant'Ana, Usama Raheel, Nurul Huda Mahmood
Submitted / Revised Submitted on 23 Jan 2026; Last revised 6 Aug 2026 (v2)
Dataset Link Kaggle Dataset Repository
Full-Text Links View PDF | TeX Source | HTML Version

Abstract

Abstract

Train delays result from complex interactions between operational, technical, and environmental factors. While weather impacts railway reliability, particularly in Nordic regions, existing datasets rarely integrate meteorological information with operational train data. This study presents the first publicly available dataset combining Finnish railway operations with synchronized meteorological observations from 2018–2024.

Train delays result from complex interactions between operational, technical, and environmental factors. While weather impacts railway reliability, particularly in Nordic regions, existing datasets rarely integrate meteorological information with operational train data. This study presents the first publicly available dataset combining Finnish railway operations with synchronized meteorological observations from 2018–2024.

The dataset integrates operational metrics from the Finland Digitraffic Railway Traffic Service with weather measurements from 209 environmental monitoring stations, using spatial-temporal alignment via Haversine distance. It encompasses 28 engineered features across operational variables and meteorological measurements, covering approximately 38.5 million observations from Finland's 5,915-kilometer rail network.

The dataset integrates operational metrics from the Finland Digitraffic Railway Traffic Service with weather measurements from 209 environmental monitoring stations, using spatial-temporal alignment via Haversine distance. It encompasses 28 engineered features across operational variables and meteorological measurements, covering approximately 38.5 million observations from Finland's 5,915-kilometer rail network.

Preprocessing includes strategic missing data handling through spatial fallback algorithms, cyclical encoding of temporal features, and robust scaling of weather data to address sensor outliers. Analysis reveals distinct seasonal patterns, with winter months exhibiting delay rates exceeding 25% and geographic clustering of high-delay corridors in central and northern Finland. Furthermore, the work demonstrates applications of the dataset in analyzing the reliability of railway traffic in Finland. A baseline experiment using XGBoost regression achieved a Mean Absolute Error of 2.73 minutes for predicting station-specific delays, demonstrating the dataset's utility for machine learning applications. The dataset enables diverse applications, including train delay prediction, weather impact assessment, and infrastructure vulnerability mapping, providing researchers with a flexible resource for machine learning applications in railway operations research.

Preprocessing includes strategic missing data handling through spatial fallback algorithms, cyclical encoding of temporal features, and robust scaling of weather data to address sensor outliers. Analysis reveals distinct seasonal patterns, with winter months exhibiting delay rates exceeding 25% and geographic clustering of high-delay corridors in central and northern Finland. Furthermore, the work demonstrates applications of the dataset in analyzing the reliability of railway traffic in Finland. A baseline experiment using XGBoost regression achieved a Mean Absolute Error of 2.73 minutes for predicting station-specific delays, demonstrating the dataset's utility for machine learning applications. The dataset enables diverse applications, including train delay prediction, weather impact assessment, and infrastructure vulnerability mapping, providing researchers with a flexible resource for machine learning applications in railway operations research.