中央研究院 資訊科學研究所

This presentation will introduce a new research method to monitor ideas, events, and trends disseminated in cyberspace (the web) and social media, called the Spatial Web Automatic Reasoning and Mapping System (SWARMS) (http://mappingideas.sdsu.edu).  We developed this innovative spatial tracking framework and utilized both commercial search engines (Bing and Yahoo) and Twitter search APIs to create visualization maps showing the density of ranked web pages or tweets related to specific keywords in different cities.  

 The Web and social media are powerful cyberspace platforms for ideas, collective thinking, and collective actions in real space. By analyzing multiple web information landscapes and spatial-selected tweets, these information maps may reveal important "geospatial fingerprints" indicating important semantic or context meanings associated with the keywords. For example, when we used the keyword, “flu”, to compare the web page densities between December 25, 2011 (623 web pages) and January 01, 2012 (617 web pages), the areas with increasing density of “flu” web pages (Dallas and Boston) are highly correlated to the weekly CDC Flu Surveillance Map on December 29, 2012.  We also found similar results with “whooping cough” keyword.

 This method was tested with the 2012 presidential candidates (Republican) in U.S. and compared them with actual primary voting results. We found some interesting spatial patterns in these maps corresponding to the voting polls.  For example, the web page density of “Mitt Romney” is much higher in Utah comparing to other Republican candidates. The maps also show a higher web page density for “Newt Gingrich” in Georgia, and a higher web page density for “Ron Paul” in southern Texas. The regional tweets analysis of Republican candidates also reflects some primary poll trends (temporal changes). 

This cyberspace mapping approach may provide a new research direction for social scientists to study human thoughts, behaviors, disease outbreaks, global web contents, and internet communication theories. The spatial-temporal analysis of web contents can help us to understand the diversity of human concepts in a global scale and to be applied in the fields of business marketing, homeland security, public policy making, and public health.

Ming-Hsiang (Ming) Tsou is Professor in the Department of Geography, San Diego State University.  He received a B.S. from National Taiwan University in 1991, an M.A. from the State University of New York at Buffalo in 1996, and a Ph.D. from the University of Colorado at Boulder in 2001, all in Geography. His research interests are in Mapping Cyberspace, Internet mapping, Web GIS, mobile GIS and wireless communication, and cyberinfrastructure with high performance computing technology. He has applied his research interests in applications such as wildfire mapping, environmental monitoring and management, habitat conservation, K-12 education, and homeland border security. He is co-author of the book, Internet GIS, published in 2003 and served on the editorial boards of the Annals of GIS (2008-) and the Professional Geographers (2011-). Tsou was the co-chair of the NASA Earth Science Enterprise Data System Working Group (ESEDWG) Standard Process Group (SPG) from 2004 to 2007 and the 2007-2008 Chair of the Cartographic Specialty Group in the Association of American Geographers (AAG). Tsou received the 2004 and 2010 Outstanding Faculty Award at San Diego State University. Tsou was appointed by the National Academy of Science in 2006 to serve on the committee on “Research Priorities for the USGS Center of Excellence for Geospatial Information Science”. In 2007, he created and maintained an interactive Web-based mapping services for San Diego Wildfires 2007 (http://map.sdsu.edu) and his efforts have been recognized by the AAG newsletters and the San Diego Union Tribune (newspaper). In 2010, Tsou served as the Principle Investigator (PI) of a NSF-CDI project," Mapping ideas from Cyberspace to Realspace", funded by National Science Foundation, Division of Computer and Network Systems, NSF Program CDI-Type II Award # 1028177. http://mappingideas.sdsu.edu/. This NSF project integrates geographic information systems (GIS), computational linguistics (CL), web search engines (Bing and Yahoo) and social media (Twitter) APIs to track and analyze public-accessible websites and tweets for visualizing web information landscapes of ideas and events in cyberspace.