Speakers: Tim Matisziw, PhD

Tim Matisziw, PhD


Associate Professor
University of Missouri


Dr. Tim Matisziw is an Associate Professor at the University of Missouri (MU).  He received his BA (1997) and MS (2000) in Geography from MU and his PhD (2005) in Geography from the Ohio State University. Since 2008 he has served as faculty at MU in the departments of Geography, Civil & Environmental Engineering, and the Institute for Data Science and Informatics. He also serves as the Director of the Geospatial Intelligence certificate programs and the Director of Graduate Studies for the Informatics PhD program at MU. Dr. Matisziw regularly teaches courses in advanced geographic information systems (GIS), transportation systems, location analysis and site selection, and the geospatial sciences in national security. His primary areas of research and expertise include spatial decision support, network modeling, transportation systems, spatial optimization, and GIS. Results emanating from this research have been published in a wide variety of journals including Computers, Environment and Urban Systems, Landscape Ecology, Geographical Analysis, Networks and Spatial Economics, Landscape and Urban Planning, International Journal of Geographical Information Science, and Computers and Operations Research. His current research is funded by the U.S. Army, the U.S. National Geospatial Intelligence Agency, The U.S. Environmental Protection Agency, and the Missouri Department of Higher Education. His professional affiliations include: M.AAG, M.ACM, M.AAAS, M.AGS, M.AGU, M.AMS, M.ASCE, M.ACM, SM.IEEE, M.RSAI, F.RGS, M.INFORMS, M.ISCB, M.SIGMAXI, and M.USGIF.

Modeling the Relationship Between the Urban Environment and Individual Sentiment

Reasoning about how elements of the urban environment may influence the sentiment of individuals is important in planning for urban design and function. To investigate the nature of these relationships, information containing indicators of individual sentiment at a particular place and time is needed as is information about ambient urban features that may influence an individual’s sentiment. In this presentation, a framework for processing and analyzing these two different types of information is described. First, social media postings are used to infer the sentiment of individuals through natural language processing. As the location and date/time of the postings are known, indicators of urban features proximate to the posting sites can then be sought. Street-level imagery can provide some insight as to the type of features an individual may perceive at a location. Provided such imagery, feature detection and semantic segmentation neural network models can be applied to classify images into features representing enclosure, complexity, and scale. The utility of the proposed framework is demonstrated using Instagram postings as indicators of individual sentiment and Google Street View imagery to infer urban features from the individual’s perspective.