| The research presented in this thesis incorporates three growing fields in geography: GIS (Geographical Information Systems), location tracking technology such as mobile GPS (Global Positioning System) and theories in transportation geography. The emergence of new technologies has provided researchers new opportunities to study people on how they progress through time and space. In the past, GPS has been used mostly to monitor vehicle movement. The focus of this thesis is on collecting person-based GPS travel data in an effort quantitatively reveal important attributes of human activity-travel patterns. In particular, this thesis presents an analysis of trips by three mode types (walk, bus and car trips) along with a model for predicting mode. A cell phone wirelessly connected to a mobile GPS acted as the primary collection device along with support technologies such as custom software and hardware. An electronic diary resides on the cell phone in which participant's record the start of each trip and activity. Diary and GPS data is matched to extract trip information. A GIS is used to produce a series of explanatory variables for each trip. The variables are based on measures of speed and proximity to the underlying road and transit network. Understanding the relationship between each mode of travel and these explanatory variables is done through the use of decision tree software. Results show that measures in speed alone do not distinguish between the three modes of transportation. The proximity to road and bus network significantly improves the decision trees model prediction for mode identification. Future avenues of research may include examining multi-modal trips and the detection of other mode types such as subway, bicycle and light rail transit. |
