README.md

Dublin2021

A Dataset for Traffic Prediction

Traffic flow data are collected from the website Transport Infrastructure Ireland Traffic Data (https://trafficdata.tii.ie/). Dataset contains traffic volume extracted from 33 traffic counters located in the center of Dublin, Ireland. The time period used is from 1/1/2021 to 31/12/2021 and the time interval is 5 minutes. This dataset can be applied to traffic prediction problems, especially using GNN-based approaches.

Data structure

Dublin2021 contains multiple data files related to spatial and temporal features of traffic flow

1. ./raw/ contains raw traffic flow data in .xls format. The structure of folder ./raw/ is as follows, with each sub-folder is data collected from a traffic counter, each folder contains 12 .xls files corresponding to each month in 2021.

   
         raw/
         |
         ├── TMU M01 000.0 N/
         │ └── 2021 (1).xls
         │ └── 2021 (2).xls
         │ └── ...
         ├── TMU M01 000.0 N/
         │ └── 2021 (1).xls
         │ └── 2021 (2).xls
         │ └── ...
         ├── ...
         

2. adj_matrix.csv is the weighted adjacency matrix of the shape [33, 33]. The value at i-th row and j-th column is the road distance from counter i to counter j.
3. counter_information.csv contains the mapping index, longitude and latitude of each traffic counter.
4. distances.csv contains information about road distance from one traffic counter to another.
5. traffic_flow.csv has 34 columns, the first one is datetime and the remaining represents traffic flow recorded at each counter every 5 minutes from 1/1/2021 to 31/12/2021.