| # Dataset information: the entry point for dataset configuration, including file locations for each modality. | |
| dataset_id: LeageST | |
| dataset_name: LargeST | |
| domain: traffic_management | |
| sample_defination: | |
| sample_type: traffic flow monitoring sensor | |
| sample_id_graph: ID2 # Corresponding column name in the data; use None if the index is used directly. | |
| sample_id_time_series: ID # Corresponding column name in the data; use None if the index is used directly. | |
| description: each sample corresponds to a traffic flow monitoring sensor | |
| modalities: | |
| # Available feature modalities. | |
| meta_features: # Static metadata, analogous to tabular data; read ca_meta.csv from the main directory. | |
| required: true # meta_features is required. | |
| file_name: ca_meta.csv # All provided files should be directly readable by pandas as a pd.DataFrame. | |
| node_level_task_graphs: | |
| available: true # Whether a graph for node-level tasks is available; if so, load its data from the corresponding directory. | |
| graph_ids: | |
| - road_distance_graph # Graph name, for example road_distance_graph (shortest driving distance graph). | |
| time_series: | |
| available: true # Whether time series data is available; if so, load it from the corresponding directory. | |
| time_series_ids: | |
| - traffic_flow # Time series name, for example traffic_flow (traffic volume). | |
| tasks: # Task construction. | |
| # Supported task types, for example meta_feature prediction, time series forecasting, and ranking. | |
| - lane_count_regression # Lane count regression. | |
| - traffic_flow_forecasting # Traffic flow forecasting. | |
| # Data source information. | |
| license: | |
| data_license: source-dependent | |
| redistribution_allowed: true | |
| source: | |
| name: Citi Bike Trip History | |
| url: source url or citation |