Traffic Prediction

This dataset contains the traffic data in San Bernardino from July to August in 2016, with 170 detectors on 8 roads with a time interval of 5 minutes. This dataset is popular as a benchmark traffic forecasting dataset.

PeMS04 is a traffic forecasting benchmark.

METR-LA is a dataset for traffic prediction.

PEMS-BAY is a dataset for traffic prediction.

PeMSD7 is traffic data in District 7 of California consisting of the traffic speed of 228 sensors while the period is from May to June in 2012 (only weekdays) with a time interval of 5 minutes. This dataset is popular for benchmark the traffic forecasting models.

In this work, we propose LargeST as a new benchmark dataset (see Figure 1), with the goal of facilitating the development of accurate and efficient methods in the context of large-scale traffic forecasting. The distinguishing characteristic of LargeST lies not only in its extensive graph size, encompassing a total of 8,600 sensors in California, but also in its substantial temporal coverage and rich node information – each sensor contains 5 years of data and comprehensive metadata. LargeST comprises four sub-datasets, each characterized by a different number of sensors. The biggest one is California (CA), including a total number of 8,600 sensors. We also construct three subsets of CA by selecting three representative areas within CA and forming the sub-datasets of Greater Los Angeles (GLA), Greater Bay Area (GBA), and San Diego (SD).

PeMS07 is a traffic forecasting benchmark.

The dataset refers to the traffic speed data in San Francisco Bay Area, containing 307 sensors on 29 roads. The time span of the dataset is January-February in 2018. It is a popular benchmark for traffic forecasting.

PeMS08 is a traffic forecasting dataset.

Taxi flow data of New York City with grid 20x10.

EXPY-TKY contains the traffic speed information and the corresponding traffic incident information in 10-minute interval for 1843 expressway road links in Tokyo over three months (2021/10∼2021/12). Compared with other benchmarks for traffic prediction, EXPY-TKY covers a larger scale and more complex incident situations. Potential tasks of EXPY-TKY include traffic prediction, incident detection, and road type classification.

Q-Traffic is a large-scale traffic prediction dataset, which consists of three sub-datasets: query sub-dataset, traffic speed sub-dataset and road network sub-dataset. Source: https://github.com/JingqingZ/BaiduTraffic Image Source: https://github.com/JingqingZ/BaiduTraffic

Bike flow data of New York City with grid 16x8.

Bike flow data of New York City.

Taxi speed data in 15min interval from 156 sensors on major roads of Luohu District in Shenzhen, China, from Jan. 1 to Jan. 31, 2015.

The original dataset from Diffusion Convolutional Recurrent Neural Network: Data-Driven Traffic Forecasting contains traffic readings collected from 207 loop detectors on highways in Los Angeles County, aggregated in 5 minutes intervals over four months between March 2012 and June 2012. The Point missing setting, introduced in Filling the G_ap_s: Multivariate Time Series Imputation by Graph Neural Networks, is a variant for imputation in which 25% of data are masked out uniformly at random. Results on this dataset are assumed to be obtained in-sample, meaning that the test interval is used also for training, excluding data used for evaluation.

The original dataset from Diffusion Convolutional Recurrent Neural Network: Data-Driven Traffic Forecasting contains 6 months of traffic readings from 01/01/2017 to 05/31/2017 collected every 5 minutes by 325 traffic sensors in San Francisco Bay Area. The measurements are provided by California Transportation Agencies (CalTrans) Performance Measurement System (PeMS). The Point missing setting, introduced in Filling the G_ap_s: Multivariate Time Series Imputation by Graph Neural Networks, is a variant for imputation in which 25% of data are masked out uniformly at random. Results on this dataset are assumed to be obtained in-sample, meaning that the test interval is used also for training, excluding data used for evaluation.

The Beijing Traffic Dataset collects traffic speeds at 5-minute granularity for 3126 roadway segments in Beijing between 2022/05/12 and 2022/07/25.

This repository contains three graph datasets for the UE traffic assignment problem on Sioux-Falls, Eastern-Massachusetts and Anaheim networks in both dgl and pyg formats. The datasets are generated and used to train and evaluate models for solving the User Equilibrium (UE) problem on three transportation networks: - ** Sioux-Falls** - ** Eastern-Massachusetts** - **Anaheim ** These networks are sourced from the well-known "transport networks for research" repository. These datasets are generated for the study "A hybrid deep-learning-metaheuristic framework for bi-level network design problems" by Bahman Madadi and Gonçalo H. de Almeida Correia, published in Expert Systems with Applications. ## Metadata Detailed information can be found in the Metadata.md file. ## Links - Figshare Dataset Repository: Link ## References - Article: A hybrid deep-learning-metaheuristic framework for bi-level network design problems - Link - GitHub Repository: HDLMF_GIN-GA - Link - Primary Figshare Data Repository: Link Note: This dataset is maintained at Figshare.

This is the static test data from the study "Global Geolocated Realtime Data of Interfleet Urban Transit Bus Iding" collected by GRD-TRT-BUF-4I. test-data-a.csv was collected from December 31, 2023 00:01:30 UTC to January 1, 2024 00:01:30 UTC. test-data-b.csv was collected from January 4, 2024 01:30:30 UTC to January 5, 2024 01:30:30 UTC. test-data-c.csv was collected from January 10, 2024 16:05:30 UTC to January 11, 2024 16:05:30 UTC.

VLUC (Video-Like Urban Computing) is a benchmark for video-like computing on citywide traffic density and crowd prediction. It consists of two new datasets BousaiTYO and BousaiOSA and existing datasets TaxiBJ, BikeNYC I-II, and TaxiNYC. Source: VLUC: An Empirical Benchmark for Video-Like Urban Computing on Citywide Crowd and Traffic Prediction

The CTV-Dataset (CTV stands for Cyclist Top-View) is a trajectories dataset for cyclist behaviour in mixed-traffic environments (aka. shared spaces). This dataset is meant to enlarge the available datasets in the community, focusing on cyclists as main road users to help the research in understanding and predicting cyclist behaviour in shared spaces. The dataset results from an experiment conducted in TU Clausthal to extract data from possible interaction scenarios with other road users, such as pedestrians and cars, in shared spaces. The scenarios were captured using a drone with 4K (3840×2160) resolution at 29.97 fps to ensure high-quality results. The trajectories were extracted using an in-house developed computer vision algorithm.