Matthew Hockenberry, Rob Gens

city, visualization, semantics, interpretation, meaning, commonsense

Can computers understand what a space means to us? We think so, and demonstrate a system that seems to feel the same way. By making use of common-sense knowledge what an average person takes a way from a place we can build visualizations that aren't dependent on what we put in a database - just what we can describe with language.

Abstract

The system begins by parsing textual descriptions of a place, what we find in a place, and what we are saying about that place. From this is builds a list of semantic tags. These tags allow us to quickly visualize meaning semantic descriptions of regions from a given building to a street, a neighborhood, and even a city.

The visualization approach focuses on building an integrated visual language for looking at the city, while allowing the deconstruction to its underlying semantic meaning.

In this presenation we will apply this technique to the City of San Jose.

URLS:
Blog and resource of the general project and research.