next generation urban emergency management: when human ... · 6. human mobility and transportation...

1
Next Generation Urban Emergency Management: When Human Emergency Mobility Modeling Meets Big Data Xuan Song and Ryosuke Shibasaki Center for Spatial Information Science, The University of Tokyo Website: http://shiba.iis.utokyo.ac.jp/song 1. Research Aims: The frequency and intensity of natural disasters has significantly increased over past decades and this trend is predicted to continue. Facing these possible and unexpected disasters, urban emergency management has become a problem for governments across the world. To plan effective humanitarian relief, transportation scheduling, disaster management and longterm societal reconstruction, modeling, understanding and accurately predicting human “emergency behavior and mobility” will play a most important role for the next generation of urban emergency management. Hence, our research aims to understand what basic laws govern human behavior and mobility following emergency events by mining big and heterogeneous data, to develop deep and powerful models for human emergency mobility prediction and simulation, and implement intelligent system for nextgeneration urban emergency management. Prediction of human emergency behavior and their mobility following largescale disaster. Can we develop some powerful and deep models to predict human emergency behavior and its movements following natural disasters? Center for Spatial Information Science, The University of Tokyo 3. Empirical Analysis of Human Behavior: 6. Human Mobility and Transportation Mode Prediction 4. Urban Mobility Modeling and People Flow Simulation Big and Heterogeneous Data. (a) shows human mobility in Tokyo when the Great East Japan Earthquake will happen, and (b) shows human mobility in Fukushima while the Fukushima Nuclear Accident happened. (c) shows the earthquake information entire Japan in March 2011. (d) shows the urban mobility graph of Fukuoka, and the edge color in indicates the edge parameters. 5. Visualization of Results 2. Big and Heterogeneous Data Human mobility data: GPS records of 1.6 million anonymized in three years. Disaster information data: 17520 times of earthquake data throughout Japan in four years. Disaster reporting data: government declarations as well as news reports from mainstream medias. Transportation data: road network data, metro network data, transportation mode label data. Traffic Accident data: traffic accident data in two years all over Japan A B C D A Tokyo B Bunkyo C Itabashi D 3. Evacuation Behaviors Discovery and Analysis 5. Human Behavior and Mobility Prediction

Upload: others

Post on 27-Jun-2020

5 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Next Generation Urban Emergency Management: When Human ... · 6. Human Mobility and Transportation Mode Prediction 4. Urban Mobility Modeling and People Flow Simulation Big and Heterogeneous

Next Generation Urban Emergency Management: When Human Emergency Mobility Modeling Meets Big Data

Xuan Song and Ryosuke Shibasaki

Center for Spatial Information Science, The University of TokyoWebsite: http://shiba.iis.u‐tokyo.ac.jp/song

1. Research Aims:The frequency and intensity of natural disasters has significantlyincreased over past decades and this trend is predicted tocontinue. Facing these possible and unexpected disasters, urbanemergency management has become a problem for governmentsacross the world. To plan effective humanitarian relief,transportation scheduling, disaster management and long‐termsocietal reconstruction, modeling, understanding and accuratelypredicting human “emergency behavior and mobility” will play amost important role for the next generation of urban emergencymanagement. Hence, our research aims to understand what basiclaws govern human behavior and mobility following emergencyevents by mining big and heterogeneous data, to develop deepand powerful models for human emergency mobility predictionand simulation, and implement intelligent system for next‐generation urban emergency management.

Prediction of human emergency behavior and their mobility following large‐scale disaster. Can we develop some powerful and deep models to predict human emergency behavior and its movements following natural disasters?

Center for Spatial Information Science, The University of Tokyo

3. Empirical Analysis of Human Behavior:

6. Human Mobility and Transportation Mode Prediction

4. Urban Mobility Modeling and People Flow Simulation

Big and Heterogeneous Data. (a) shows human mobility in Tokyo when the Great East JapanEarthquake will happen, and (b) shows human mobility in Fukushima while the Fukushima NuclearAccident happened. (c) shows the earthquake information entire Japan in March 2011. (d) shows theurban mobility graph of Fukuoka, and the edge color in indicates the edge parameters.

5. Visualization of Results

2. Big and Heterogeneous Data Human mobility data: GPS records of 1.6 million anonymized in three years. Disaster information data: 17520 times of earthquake data throughout 

Japan in four years. Disaster reporting data: government declarations as well as news reports 

from mainstream medias. Transportation data: road network data, metro network data, 

transportation mode label data.  Traffic Accident data: traffic accident data in two years all over Japan

A B C D

A TokyoB BunkyoC ItabashiD

3. Evacuation Behaviors Discovery and Analysis 5. Human Behavior and Mobility Prediction