Traffic flow prediction datasets

HZMetro  : The HZMetro dataset encompasses data collected from 80 subway stations spanning from January 1st, 2019, to January 25th, 2019. It presents daily information from 05:30 to 23:30, consolidated into 15-minute intervals, resulting in 73 time steps per day. This dataset records the total count of individuals entering and exiting each station within these 15-minute intervals.

SHMetro : The SHMetro dataset comprises data gath ered from 288 subway stations over the period from July 1st, 2016, to September 30th, 2016. It presents daily information from 05:30 to 23:30, and organized into 15-minute intervals, amounting to 73 time steps per day. Similar to the HZMetro dataset, it documents the total count of individuals entering and exiting each station within these 15-minute intervals.

PEMS08: The PEMS08 dataset contains traffic data from a highway network spanning from July 1, 2016, to July 31, 2016. Aggregated at 5-minute intervals, it comprises 288 intervals per day. This dataset is gathered from 170 sensors located on the highway.

WHBT: The WHBT dataset is composed of traffic data from various bridges and tunnels in a certain city, from June 1, 2024 to July 5, 2024. The city has a total of 780 bridges and tunnels, including 502 bridges, 216 flyovers, and 62 tunnels. The distribution of these bridges and tunnels is shown in Fig.6(a). According to their actual road network connection conditions, based on the real distance between each bridge and tunnel and the time required for passage, a computational graph of the city’s bridges and tunnels network is constructed, as shown in Fig.6(b). Among them, 251 bridges and tunnels have traffic speed data with a time interval of five minutes. The computation graph generated from the complete bridge and tunnel network for these 251 bridges and tunnels is shown in Fig.7.