A Large Scale PeMS Traffic Speed Dataset

<p>This repository is a traffic speed data set collected by California Freeway Performance Measurement System (PeMS). In this data set, it has 11160 time series corresponding to <strong>11160</strong> sensors/locations. In three data files (i.e., <em>pems-4w.csv</em>, <em>pems-8w.csv</em>, and <em>pems-12w.csv</em>), each time series is indeed a time series with <strong>288</strong> (5-minute) time points per day during the first <strong>4 weeks</strong> (PeMS-4W), first <strong>8 weeks</strong> (PeMS-8W), and first <strong>12 weeks</strong> (PeMS-12W) of 2018, respectively. Note that the original data was downloaded from PeMS system with 30-second traffic speed records. Our data is a subset of file&nbsp;<em>speed.h5</em>&nbsp;created by article <a href="https://arxiv.org/pdf/2004.08038.pdf">Transfer Learning with Graph Neural Networks for Short-Term Highway Traffic Forecasting</a>&nbsp;and its&nbsp;original link is available at&nbsp;<a href="https://bit.ly/2SsGleK">here</a>.</p> <p>Our research work on this data set is:</p> <p>Xinyu Chen, Yixian Chen, Nicolas Saunier, Lijun Sun (2021). Scalable low-rank tensor learning for spatiotemporal traffic data imputation. Transportation Research Part C: Emerging Technologies, 129: 103226.&nbsp; [<a href="https://doi.org/10.1016/j.trc.2021.103226">DOI</a>] [<a href="https://arxiv.org/abs/2008.03194">preprint</a>] [<a href="https://doi.org/10.5281/zenodo.3939792">data</a>] [<a href="http://github.com/xinychen/tensor-learning">Python code</a>]</p>