CityNet: A Multi-city Multi-modal Dataset for Smart City Applications

Data-driven approaches have been applied to many problems in urban computing. However, in the research community, such approaches are commonly studied under data from limited sources, and are thus unable to characterize the complexity of urban data coming from multiple entities and the correlations among them. Consequently, an inclusive and multifaceted dataset is necessary to facilitate more extensive studies on urban computing. In this paper, we present CityNet, a multimodal urban dataset containing data from 7 cities, each of which coming from 3 data sources. We first present the generation process of CityNet as well as its basic properties. In addition, to facilitate the use of CityNet, we carry out extensive machine learning experiments, including spatio-temporal predictions, transfer learning, and reinforcement learning. The experimental results not only provide benchmarks for a wide range of tasks and methods, but also uncover internal correlations among cities and tasks within CityNet that, with adequate leverage, can improve performances on various tasks. With the benchmarking results and the correlations uncovered, we believe that CityNet can contribute to the field of urban computing by supporting research on many advanced topics.

[1]  Jieping Ye,et al.  Spatiotemporal Multi-Graph Convolution Network for Ride-Hailing Demand Forecasting , 2019, AAAI.

[2]  Guangzhong Sun,et al.  Driving with knowledge from the physical world , 2011, KDD.

[3]  Pravin Varaiya,et al.  Freeway performance measurement system (pems) , 2002 .

[4]  Max Welling,et al.  Semi-Supervised Classification with Graph Convolutional Networks , 2016, ICLR.

[5]  Qiang Yang,et al.  A Survey on Transfer Learning , 2010, IEEE Transactions on Knowledge and Data Engineering.

[6]  Zhe Xu,et al.  Large-Scale Order Dispatch in On-Demand Ride-Hailing Platforms: A Learning and Planning Approach , 2018, KDD.

[7]  Qiang Yang,et al.  Transfer Knowledge between Cities , 2016, KDD.

[8]  Yu Zheng,et al.  AutoST: Efficient Neural Architecture Search for Spatio-Temporal Prediction , 2020, KDD.

[9]  Zhe Xu,et al.  A Deep Value-network Based Approach for Multi-Driver Order Dispatching , 2019, KDD.

[10]  Jieping Ye,et al.  Multi-Agent Reinforcement Learning for Order-dispatching via Order-Vehicle Distribution Matching , 2019, CIKM.

[11]  Xianfeng Tang,et al.  Learning from Multiple Cities: A Meta-Learning Approach for Spatial-Temporal Prediction , 2019, WWW.

[12]  Cyrus Shahabi,et al.  Diffusion Convolutional Recurrent Neural Network: Data-Driven Traffic Forecasting , 2017, ICLR.

[13]  Xiao-Ming Wu,et al.  Deeper Insights into Graph Convolutional Networks for Semi-Supervised Learning , 2018, AAAI.

[14]  Tianjian Chen,et al.  Federated Machine Learning: Concept and Applications , 2019 .

[15]  Yu Zheng,et al.  Deep Spatio-Temporal Residual Networks for Citywide Crowd Flows Prediction , 2016, AAAI.

[16]  Jun Wang,et al.  Efficient Ridesharing Order Dispatching with Mean Field Multi-Agent Reinforcement Learning , 2019, WWW.

[17]  Xing Xie,et al.  Discovering spatio-temporal causal interactions in traffic data streams , 2011, KDD.

[18]  Jignesh M. Patel,et al.  Big data and its technical challenges , 2014, CACM.

[19]  Yike Guo,et al.  Deep Sequence Learning with Auxiliary Information for Traffic Prediction , 2018, KDD.

[20]  Cláudio T. Silva,et al.  Visual Exploration of Big Spatio-Temporal Urban Data: A Study of New York City Taxi Trips , 2013, IEEE Transactions on Visualization and Computer Graphics.

[21]  Zhanxing Zhu,et al.  Spatio-temporal Graph Convolutional Neural Network: A Deep Learning Framework for Traffic Forecasting , 2017, IJCAI.

[22]  Xiuwen Yi,et al.  DNN-based prediction model for spatio-temporal data , 2016, SIGSPATIAL/GIS.

[23]  Feng Liu,et al.  Cross-City Transfer Learning for Deep Spatio-Temporal Prediction , 2018, IJCAI.

[24]  Licia Capra,et al.  Urban Computing: Concepts, Methodologies, and Applications , 2014, TIST.