A Collection of resources for traffic prediction with external factors apart from normal spatial (Network structure, connectivity) and temporal (Time-of-day, Day-in-week) features.
- Types of External Factor:
π§οΈ - Meteorology, including weather conditions, temperature, wind speed etc
π - Holidays or other temporal features
π§ - Events, including traffic accidents, traffic conditions (congestion), social events and public activities
πͺ - POIs or Land features
π£οΈ - Road or highway characteristics, e.g. link width, link length/distance, link type, number of lane, having bus stops, etc
- Methods of Integrating External Factors
β - Concatenation
π - LSTM, GRU
π₯ - Fusion
π’ - Knowledge representation learning
| Journal/ Conference |
Title | Model | Factors | Code | Data |
|---|---|---|---|---|---|
| SIGSPATIAL '15 | Traffic prediction in a bike-sharing system Paper Page | based on GBRT | π§οΈ | Code | NYC, D.C. Bike |
| IEEE TVT | Improving traffic flow prediction with weather information in connected cars: A deep learning approach Paper | based on DBNs, MTL | π§οΈ | PeMS MesoWest |
|
| IEEE TKDE | Citywide Traffic Volume Estimation Using Trajectory Data Paper | π§οΈπͺ | Beijing | ||
| AAAI '17 | Deep Spatio-Temporal Residual Networks for Citywide Crowd Flows Prediction Paper Page | ST-ResNet | π§οΈπ | ST-ResNet-Pytorch STResNet-PyTorch ST-ResNet STResNet |
Beijing NYC |
| SIGSPATIAL '18 | Bike Flow Prediction with Multi-Graph Convolutional Networks Paper | Multi-Graph Convolutional Neural Network | π§οΈ | GraphCNN-Bike | NYC bike CHI Bike NOAA NCEI |
| KDD '18 | Deep sequence learning with auxiliary information for traffic prediction Paper | based on Seq2Seq+ | π§ | BaiduTraffic | Q-Traffic |
| IJCAI-18 | GeoMAN: Multi-level Attention Networks for Geo-sensory Time Series Prediction Paper | GeoMAN | π§οΈπͺ | GeoMAN(Tensorflow) GeoMAN(Pytorch) |
Water quality Air quality |
| AAAI '18 | Deep multi-view spatial-temporal network for taxi demand prediction Paper | DMVST-Net | π§οΈπ | DMVST-Net | DiDi(Guangzhou) |
| IET ITS | Combining weather condition data to predict traffic flow: A GRUβbased deep learning approach Paper | DGRNN | π§οΈ | PeMS NOAA |
|
| IEEE MDM '19 | Temporal Graph Convolutional Networks for Traffic Speed Prediction Considering External Factors Paper | GTCN | π§οΈπ πͺ | PEMSD7,PEMSD4 | |
| AAAI '19 | Spatiotemporal Multi-Graph Convolution Network for Ride-Hailing Demand Forecasting Paper | ST-MGCN | πͺ | ST-MGCN | Beijing Shanghai |
| AAAI '19 | DeepSTN+: Context-Aware Spatial-Temporal Neural Network for Crowd Flow Prediction in Metropolis Paper | DeepSTN+ | πͺ | DeepSTN | MobileBJ BikeNYC |
| AAAI '19 | Revisiting Spatial-Temporal Similarity: A Deep Learning Framework for Traffic Prediction Paper | STDN | π§οΈπ§(But not use in experiment) | STDN | NYC-Taxi NYC-Bike |
| KDD '19 | UrbanFM: Inferring Fine-Grained Urban Flows Paper | UrbanFM | π§οΈπ π§ | UrbanFM | TaxiBJ HappyValley |
| KDD '19 | Urban traffic prediction from spatio-temporal data using deep meta learning Paper | ST-MetaNet | πͺπ£οΈ | ST-MetaNet | TDrive(flow) METR-LA(speed) |
| Journal/ Conference |
Title | Model | Factors | Code | Data |
|---|---|---|---|---|---|
| AAAI '20 | Multi-Range Attentive Bicomponent Graph Convolutional Network for Traffic Forecasting Paper | MRA-BGCN | π£οΈ | MAR-BGCN_GPU_pytorch | METR-LA PEMS-BAY |
| IET ITS | Multiβstep traffic speed prediction model with auxiliary features on urban road networks and its understanding Paper | EGC-LSTM | π§οΈπ π£οΈ | Guiyang NOAA |
|
| TR_C | Evaluation and prediction of transportation resilience under extreme weather events: A diffusion graph convolutional approach Paper | based on DCRNN and dynamic-capturing algorithm | π§οΈ | resilience_shenzhen | DiDi NCAR |
| IEEE TKDE | Flow prediction in spatio-temporal networks based on multitask deep learning Paper | MDL | π§οΈπ | MDL(uno) | NYC Taxi |
| IEEE TITS | DeepSTD: Mining Spatio-Temporal Disturbances of Multiple Context Factors for Citywide Traffic Flow Prediction Paper | DeepSTD | π§οΈπ πͺ | DeepSTD | Xiamen DiDi(Chengdu) |
| CIKM '20 | Deep Graph Convolutional Networks for Incident-Driven Traffic Speed Prediction Paper | DIGC-Net | π§οΈπ§ | SFO NYC Yahoo Weather API |
| Journal/ Conference |
Title | Model | Factors | Code | Data |
|---|---|---|---|---|---|
| IEEE TITS | STNN: A Spatio-Temporal Neural Network for Traffic Predictions Paper | STNN | πͺπ£οΈ | HK-KL, ST, TM | |
| Applied Intelligence | A multi-mode traffic flow prediction method with clustering based attention convolution LSTM Paper | CACLSTM | π§οΈπ | TaxiBJ | |
| Wiley JAT | The Prediction of Multistep Traffic Flow Based on AST-GCN-LSTM Paper | AST-GCN-LSTM | π§οΈπͺ | Luohu(Shenzhen) | |
| WWW '21 | Fine-Grained Urban Flow Prediction Paper | STRN | π§οΈπ π§πͺ | TaxiBJ HappyValley |
|
| IEEE TITS | Dynamic Spatial-Temporal Representation Learning for Traffic Flow Prediction Paper | ATFM | π§οΈπ | ATFM | TaxiBJ BikeNYC TaxiNYC |
| Elsevier EAAI | A flexible deep learning-aware framework for travel time prediction considering traffic event Paper | MC-Net | π§ | Beijing (from Amap) | |
| IEEE Access | Hybrid Deep Spatio-Temporal Models for Traffic Flow Prediction on Holidays and Under Adverse Weather Paper | CL-CN-G CL-CNG G-CN-CL |
π§οΈπ | PeMS MesoWest |
|
| IEEE Access | AST-GCN: Attribute-augmented spatiotemporal graph convolutional network for traffic forecasting Paper | AST-GCN | π§οΈπͺ | AST-GCN | SZ_taxi |
| Journal/ Conference |
Title | Model | Factors | Code | Data |
|---|---|---|---|---|---|
| Elsevier NN | Exploiting dynamic spatio-temporal graph convolutional neural networks for citywide traffic flows prediction Paper | GCN-DHSTNet | π§οΈπ π§ | TaxiBJ BikeNYC |
|
| Information Sciences | Spatialβtemporal short-term traffic flow prediction model based on dynamical-learning graph convolution mechanism Paper | Loc-GCLSTM | π§οΈ | OpenITS METR-LA |
|
| Transportmetrica B | DLW-Net model for traffic flow prediction under adverse weather Paper | DLW-Net | π§οΈ | Hefei(OpenITS) Sacramento(PeMS) MesoWest |
|
| IEEE TKDE | Predicting Citywide Crowd Flows in Irregular Regions Using Multi-View Graph Convolutional Networks Paper | MVGCN | πͺπ£οΈ | TaxiNYC TaxiBJ BikeDC BikeNYC |
|
| KeAi DCAN | Attention-based spatio-temporal graph convolutional network considering external factors for multi-step traffic flow prediction Paper | ABSTGCN-EF | π π§ | PeMS-LA PeMS-Bay |
|
| WSDM '22 | ST-GSP: Spatial-temporal global semantic representation learning for urban flow prediction Paper | ST-GSP | π§οΈπ | STGSP | TaxiBJ BikeNYC |
| IEEE TITS | KST-GCN: A Knowledge-Driven Spatial-Temporal Graph Convolutional Network for Traffic Forecasting Paper | KST-GCN | π§οΈπͺ | KST-GCN | Luohu(Shenzhen) |
| Wiley JAT | Metro Traffic Flow Prediction via Knowledge Graph and Spatiotemporal Graph Neural Network Paper | KGR-STGNN | π§οΈπ§ | BJMF15(Beijing) |
| Journal/ Conference |
Title | Model | Factors | Code | Data |
|---|---|---|---|---|---|
| IEEE ICCCNT '23 | Traffic Prediction Using Auxiliary Information Based on Mlbsae-A Hybrid Deep Learning Framework Paper | MLBSAE | π§οΈπ π§ | PeMS | |
| IEEE ISPDS '23 | An urban traffic flow prediction approach integrating external factors based on deep learning and knowledge graph Paper | KR-EAR | π§οΈπͺ | Luohu(Shenzhen) | |
| AAAI '23 | Spatio-Temporal Meta-Graph Learning for Traffic Forecasting Paper | MegaCRN | π§(Used for robustness analysis) | MegaCRN | METR-LA PEMS-BAY EXPY-TKY |
| Elsevier ESWA | Spatio-temporal graph mixformer for traffic forecasting Paper | STGM | π§οΈ(But not use in experiment) | stgm | PeMS-Bay PeMSD7M METR-LA |
| IEEE TKDE | Forecasting Fine-Grained Urban Flows Via Spatio-Temporal Contrastive Self-Supervision Paper | UrbanSTC | π§οΈπ | UrbanSTC | TaxiBJ BikeNYC |
| ECAI 2023 | WeaGAN: Weather-aware graph attention network for traffic prediction Paper | WeaGAN | π§οΈ | WeaGAN | PeMS OpenWeather |
| Physica A | Traffic flow prediction under multiple adverse weather based on self-attention mechanism and deep learning models Paper | DHA | π§οΈ | PeMS MesoWest |
|
| Elsevier KBS | Spatio-temporal fusion and contrastive learning for urban flow prediction Paper | ST-FCL | π§οΈπ | TaxiBJ BikeNYC |
|
| Spatio-temporal graph convolution network based on multimodal feature fusion (εΊδΊε€ζ¨‘ζηΉεΎθεηζΆη©ΊεΎε·η§―η½η») Paper | AFFGCN | π§οΈ | AFFGCN | PEMSD4 PEMSD8 Iowa Atmospheric Observatory |
|
| CIKM '23 | Mask- and Contrast-Enhanced Spatio-Temporal Learning for Urban Flow Prediction Paper | MC-STL | π§οΈπ | MCSTL | TaxiBJ BikeNYC |
| Nature CIS | Integrating knowledge representation into traffic prediction: a spatialβtemporal graph neural network with adaptive fusion features Paper | KR-STGNN | π§οΈπͺ | SZ-taxi SZ_POI SZ_Weather |
|
| MDPI Electronics | HIT-GCN: Spatial-Temporal Graph Convolutional Network Embedded with Heterogeneous Information of Road Network for Traffic Forecasting Paper | HIT-GCN | π§οΈπͺ | SZ-taxi SZ_POI SZ_Weather |
|
| IEEE TITS | A Deep Learning Approach for Long-Term Traffic Flow Prediction With Multifactor Fusion Using Spatiotemporal Graph Convolutional Network Paper | π§οΈ | Minnesota NOAA |
||
| IEEE TVT | Unified Spatial-Temporal Neighbor Attention Network for Dynamic Traffic Prediction Paper | USTAN | π§οΈπͺ | PriTra TaxiBJ METR-LA PEMS-BAY POI Waether Event |
| Journal/ Conference |
Title | Model | Factors | Code | Data |
|---|---|---|---|---|---|
| IET ITS | Combining weather factors to predict traffic flow: A spatialβtemporal fusion graph convolutional networkβbased deep learning approach Paper | STFGCN | π§οΈ | PEMSD4 MesoWest |
|
| Information Sciences | A multi-channel spatial-temporal transformer model for traffic flow forecasting Paper | MC-STTM | π§οΈπ§(But not use due to limited information) | PEMS03 PEMS04 PEMS07 PEMS08 METR-LA PEMS-BAY |
|
| Scientific Reports | A multiβfeature spatialβtemporal fusion network for traffic flow prediction Paper | ATFEM | π§οΈπ | Guizhou | |
| Elsevier EAAI | A traffic-weather generative adversarial network for traffic flow prediction for road networks under bad weather Paper | TWeather-GAN | π§οΈ | PeMS MesoWest |
|
| T&F SSCE | Short-term traffic flow prediction at isolated intersections based on parallel multi-task learning Paper | MTL-fusion | π | Nanhu(Jiaxing) | |
| ACM CIKM '24 | Empowering traffic speed prediction with auxiliary feature-aided dependency learning Paper | ARIAN | π§οΈπ | PeMS-D4 PeMS-D8 Cityway |
|
| ACM CIKM '24 | Seeing the Forest for the Trees: Road-Level Insights Assisted Lane-Level Traffic Prediction Paper | McgVAE | π£οΈ | McgVAE | PeMS(_F) HuaNan(Guangzhou) |
| Elsevier AI | Dual-track spatio-temporal learning for urban flow prediction with adaptive normalization Paper | DualST | π§οΈπ | TaxiBJ BikeNYC |
|
| ACM TKDE | DeepMeshCity: A Deep Learning Model for Urban Grid Prediction Paper | DeepMeshCity | π§οΈπ | DeepMeshCity | TaxiBJ BousaiTYO BousaiOSA |
| Journal/ Conference |
Title | Model | Factors | Code | Data |
|---|---|---|---|---|---|
| Scientific Reports | Traffic flow prediction based on spatial-temporal multi factor fusion graph convolutional networks Paper | STFGCN | π§οΈ | PEMS03 PEMS04 PEMS07 PEMS08 |
|
| Scientific Reports | An urban road traffic flow prediction method based on multi-information fusion Paper | MIFPN | π§οΈπ πͺ | SZ-taxi SZ_POI SZ_Weather |
|
| Scientific Reports | Linear attention based spatiotemporal multi graph GCN for traffic flow prediction Paper | LASTGCN | π§οΈπ | PEMSD4,PEMSD8 NOAA |
|
| Nature NCAA | ASTGCN for traffic flow prediction based on weather influence Paper | WI-ASTGCN | π§οΈ | PEMSD4,PEMSD8 MesoWest |
|
| Applied Soft Computing | MGC-DMF: A traffic flow forecasting method based on multi-graph spatio-temporal convolution and dynamic metric fusion with multi-source basic information Paper | MGC-DMF | π§οΈ(Used for predictive performance stability analysis) | MGC-DMF | PEMSD4 PEMSD8 SZ-taxi |
| T&F TPT | Short-term traffic flow prediction based on adaptive graph convolutional recurrent network under multi-factor fusion Paper | EAG-LSTMA | π§οΈπ π§ | Longgang(SZ) | |
| IEEE TVT | Condition-Guided Urban Traffic Co-Prediction With Multiple Sparse Surveillance Data Paper | CSTN | πͺπ£οΈ | NYC taxi NYC Bike NYC Accident External factors |
|
| KDD '25 | CausalMob: Causal Human Mobility Prediction with LLMs-derived Human Intentions toward Public Events Paper | CausalMob | π§ | CausalMob | Japan |
| Elsevier NN | Enhancing urban flow prediction via mutual reinforcement with multi-scale regional information Paper | MR-UFP | π§οΈπ πͺ | Data-of-MR-UPF | TaxiBJ BikeNYC Region category |
DiDi: Open datasets from Didi Chuxing GAIA Initiative, tracing the records of taxi drivers and passengers in cities, China. It seems it is not applicable now.
Commonly used public datasets summary