Integrated Urban Traffic-Flood (IUTF) Dataset

OverviewThe Integrated Urban Traffic-Flood (IUTF) Dataset is a comprehensive collection of urban traffic and environmental data from 16 diverse cities across<b> Europe, North America, and Asia</b>. These cities include <b>Augsburg, Cagliari, Darmstadt, Essen, Hamburg, Innsbruck, London, Lucerne, Madrid, Manchester, Marseille, Paris, Strasbourg, Taipei, Turin, and Toronto</b>. This dataset uniquely combines traffic flow information, road network data, and rainfall data to provide a robust foundation for studying urban traffic dynamics under various weather conditions, particularly during flood events.Data DescriptionFor each city, the dataset includes the following files:{city}_data_hours.npz: Traffic flow data based on the road network, containing attributes for flow, occupancy, and speed.{city}_distance_hours.csv: Spatial relationship data of the traffic network, with attributes for 'from' node, 'to' node, and distance.{city}_sensor.csv: Sensor data specific to each city.detectors_public.csv: Spatial location data for all traffic sensors.links.csv: Data linking sensors to their respective road network segments.rainfall_data.csv: Rainfall data corresponding to the time periods of sensor measurements.roads.gpkg: Road network data for the area covered by the sensors. Some cities may have multiple .gpkg files if the road network spans multiple regions. These files should be merged for comprehensive analysis. The data is sourced from OpenStreetMap.selected_network_4326.geojson: Road centreline data for the area covered by the sensors.The IUTF Dataset addresses common challenges in urban traffic-flood studies by integrating diverse data types. It offers a unique resource for researchers and practitioners in urban planning, traffic management, and climate resilience. The dataset's innovative features include the transformation of point-based traffic data to road segment attributes and the use of a line-graph topology, providing new possibilities for analysing and modelling complex urban systems. This dataset not only supports the development of advanced traffic prediction models but also facilitates research in urban resilience and traffic management during extreme weather events. It provides a more accurate representation of traffic dynamics and their interaction with environmental factors, which is crucial for developing effective strategies for urban flood resilience.Data SourceThe city traffic flow data in IUTF is from UTD19. UTD19[1] is another significant dataset used in this research, which includes urban traffic data from 40 cities worldwide. The dataset, as described in the UTD19 manual, contains detailed traffic measurements collected from various stationary sensors such as inductive loop detectors, supersonic detectors, cameras, and Bluetooth detectors. These sensors provide data on fundamental traffic variables including flow, speed, and occupancy. However, the dataset does not inherently include the spatial relationships between sensors. To overcome this, we used OSMNX[2] to retrieve OpenStreetMap (OSM) data to map the sensor locations onto the road network. By associating each sensor with its corresponding road segment, we were able to construct a graph network that accurately reflects the spatial relationships between sensors, thus enabling more detailed and context-aware traffic analysis. In addition, weather Data for London is also incorporated into the study to account for environmental factors that might affect traffic flow. This data is sourced from the London Met Office[3] and NW3 weather[4], providing detailed meteorological information such as temperature, precipitation, and wind speed. These variables are crucial for understanding and predicting traffic patterns under varying weather conditions.ReferenceLoder, A., Ambühl, L., Menendez, M. &amp; Axhausen, K. W. Understanding traffic capacity of urban networks. Sci. Rep. 9, 16283 (2019).Boeing, G. Modeling and Analyzing Urban Networks and Amenities with OSMnx.Weather and climate change. Met Office https://www.metoffice.gov.uk/ (2024).Rodgers, B. NW3 Weather - Live and historical weather from Hampstead, London. http://nw3weather.co.uk/.

### 概述
城市综合交通-洪涝（Integrated Urban Traffic-Flood, IUTF）数据集是一套涵盖欧洲、北美及亚洲共16座多元城市的城市交通与环境数据综合集。涉及城市包括奥格斯堡、卡利亚里、达姆施塔特、埃森、汉堡、因斯布鲁克、伦敦、卢塞恩、马德里、曼彻斯特、马赛、巴黎、斯特拉斯堡、台北、都灵以及多伦多。本数据集创新性地融合了交通流信息、道路网络数据与降雨数据，可为探究不同天气条件下（尤其是洪涝事件期间）的城市交通动态提供坚实的研究基础。

## 数据说明
针对每座城市，本数据集包含如下文件：
1. `{city}_data_hours.npz`：基于道路网络的交通流数据，包含流量、占有率与速度三类属性。
2. `{city}_distance_hours.csv`：交通网络空间关系数据，包含「起始节点」「终止节点」与「距离」三类属性。
3. `{city}_sensor.csv`：各城市专属的传感器数据。
4. `detectors_public.csv`：所有交通传感器的空间位置数据。
5. `links.csv`：关联传感器与其对应道路网络路段的数据。
6. `rainfall_data.csv`：与传感器测量时段相对应的降雨数据。
7. `roads.gpkg`：传感器覆盖区域的道路网络数据。若某城市的道路网络跨越多个行政区域，则可能存在多个`.gpkg`文件，需将这些文件合并后才可开展全面分析。本数据源自开放街道地图（OpenStreetMap, OSM）。
8. `selected_network_4326.geojson`：传感器覆盖区域的道路中心线数据。

本数据集通过融合多源异构数据，解决了城市交通-洪涝研究中的常见难题，可为城市规划、交通管理与气候韧性领域的研究者与从业者提供独特的研究资源。其创新特性包括将基于点位的交通数据转换为路段属性，以及采用线图拓扑结构，为复杂城市系统的分析与建模提供了全新可能。本数据集不仅可为先进交通预测模型的开发提供支撑，还可助力极端天气事件下的城市韧性与交通管理研究。其能够更精准地刻画交通动态及其与环境要素的交互关系，这对于制定行之有效的城市洪涝韧性策略至关重要。

## 数据来源
本数据集的城市交通流数据源自UTD19。UTD19[1]是本研究使用的另一套重要数据集，涵盖了全球40座城市的城市交通数据。根据UTD19手册中的描述，该数据集包含从各类固定式传感器（如感应线圈检测器、超声波检测器、摄像头与蓝牙检测器）采集的精细化交通测量数据，可提供流量、速度与占有率等核心交通变量。但该数据集本身并未包含传感器间的空间关联关系。为解决这一问题，我们借助OSMNX[2]获取开放街道地图（OpenStreetMap, OSM）数据，将传感器位置映射至道路网络中。通过将每个传感器与其对应的道路路段进行关联，我们得以构建能够精准反映传感器间空间关系的图网络，从而支持更精细化且具备上下文感知能力的交通分析。

此外，本研究还纳入了伦敦的气象数据，以考量可能影响交通流的环境要素。该数据源自伦敦气象局[3]与NW3 Weather[4]，可提供温度、降水量与风速等精细化气象变量。这些变量对于理解与预测不同天气条件下的交通模式至关重要。

## 参考文献
1. Loder, A., Ambühl, L., Menendez, M. & Axhausen, K. W. 解析城市网络的交通通行能力. 科学报告, 9, 16283 (2019).
2. Boeing, G. 借助OSMNx对城市网络与公共设施进行建模与分析. 气象与气候变化. 英国气象局 https://www.metoffice.gov.uk/ (2024).
3. Rodgers, B. NW3 Weather——来自伦敦汉普斯特德的实时与历史气象数据. http://nw3weather.co.uk/.