Detecting Activity Types and Trip Purposes from Passive GPS Data: A Data Mining Approach - Space-Time Integration in Geography and GIScience - page 206

Chapter 13
Detecting Activity Types and Trip Purposes
from Passive GPS Data: A Data Mining
Approach

Donggen Wang and Bingxia Sun

13.1 Introduction

The increasing demand for more accurate and more detailed data for activity-travel
behavior studies has posted great challenges to the conventional data collection
methods which suffer from under-reporting, inaccurate information on time and
location and low response rate. The development of positioning technologies such
as Global Positioning System (GPS) and the integration of GPS with ground-based
wireless communication network have provided new opportunities for collecting
activity-travel behavior data (Wolf 2000 ). Positioning technologies offer the possi-
bility to trace an individual's spatiotemporal trajectory - .i.e., recording the locations
of individuals second by second and the speed of movement between locations. As
individuals' activity and time use patterns are imbedded in their daily spatiotemporal
trajectories, since the 1990s, many have proposed to use GPS to collect activity and
time use data (Timmermans et al. 2009 ;Chenetal. 2010 ). It is anticipated that
the GPS-based method can not only collect behavior data with better accuracy and
completeness, but also reduce the burden on respondents greatly (Draijer et al. 2000 ;
Forrest and Pearson 2005 ; Stopher et al. 2006 ).
There are two major approaches of using GPS for collecting activity-travel
behavior data: interactive GPS and passive GPS. The interactive approach combines
GPS with the Internet or other methods and requests respondents to supplement or
verify information collected by GPS. For examples, Stopher and Collins ( 2005 ) sug-
gest using GPS together with Internet-based prompted recall survey to supplement
each other in the collection of trip-making information. Bohte and Maat ( 2008 )

;