Reflecting Human Knowledge of Place and Route-Choice Behavior Using Big Data

Exploring human knowledge of geographical space and related behavior not only helps in understanding human-environment interactions and dynamic geographic processes, but also advances Geographic Information Systems (GIS) toward a human-centric paradigm to make daily life more efficient. Today’s relatively easy acquisition of various big data provides an unprecedented opportunity for geographers to answer research questions that previously could not be adequately addressed. However, new challenges also arise regarding data quality and bias as well as change in methodology for dealing with big data that are different from traditional data types.

Representing people’s perception of place and studying driver’s route-choice behavior are two of the many applications of big data in answering research questions about human knowledge and behavior in the fields of GIS and transportation. Incorporating three papers, this dissertation focuses on these two different applications to achieve the following objectives: 1) examine the degree to which a geographic place’s spatial extent can be estimated from human-generated geotagged photos; 2) address the challenge of geotagged photos’ uneven spatial distribution in place estimation and explore an approach that can better derive a place’s spatial extent; 3) develop a method that can properly estimate the spatial extent of a place that has multiple disjoint regions while considering geotagged photos’ uneven distribution; 4) explore useful spatiotemporal patterns of taxi drivers’ route-choice behavior in a dynamic urban environment.

This dissertation makes three major contributions to big data applications’ systematic theory: 1) proposes an effective approach to handling the uneven spatial distribution problem of geotagged photos as a type of volunteered geographic data by modeling their representativeness; 2) develops methods that can properly derive the vague spatial extent of a place with or without disjoint regions; and 3) explores taxi drivers’ route-choice patterns in different situations that can inform future transportation decisions and policy-making processes.