GPS Data and Car Drivers’ Parking Search Behavior in the City of Turnhout, Belgium - Geoinformatics for Intelligent Transportation - page 247

Table 3 Parameter estimates of regression analyses
Attributes Parameters Signi cance

Constant − 0.297 0.100

Walking distance between street segment and center 0.047 0.001

Distance between street segment and closest parking facility 0.010 0.000

−

Presence of shops 0.365 0.039

Parking tariff 0.649 0.000

Goodness-of-fit
F-value 25.538 (sign. 0.000)

R-square 0.163

Adjusted R-square 0.157

The performance of the model is poor (based on adjusted R-square value) but still
acceptable (based on F-value).
The parameters show the contribution of attributes to the number of car drivers
using the street segment. A positive sign means that an increase of the attribute levels
results in a higher number of searchers in the street segment, while a negative sign
indicates a decrease of searchers when the attribute level increases. The positive sign
for the distance indicates that car drivers start to search for a free parking space at
some distance from the center (see before). If a car driver approaches a parking
facility, he/she stops searching and drives immediately to the parking facility. Streets
close to parking facilities are less used for searching. In streets with shops the
number of searchers is higher than in streets without shops. Finally, the parameter
estimate of parking tariffs shows that the higher the parking tariff the higher the
number of searchers in a street.

5 Conclusions

The study described in this paper aims to provide more insight into the temporal
and spatial aspects of car drivers

'

parking search behavior in central business
districts and shopping areas. Special attention is paid to the collection of empirical
data using GPS loggers. It appears that GPS data could be used for describe car
drivers search behavior both in terms of time and location. The approach provides

information regarding search time which can be included in accessibility studies.
Also detailed information becomes available regarding the use of street segments
that could be used in livability studies. Young [ 7 ] summarizes this as follows:
Parking search models provide an ability to investigate long-term commitments to
parking expenditure, the impact of parking information on route choice, the time
spent in searching for a space, and the choice strategy.
The temporal and spatial data could be explored in more detail when more data
(more trips and more car drivers) were available. The same holds for the modeling

'