The stay coefficient: a novel quantification of the relationship between stay time and travel time for urban shopping behavior analysis

Figures & data

Table 1. Summary of survey data.

Survey data	Tokyo metropolitan area PT survey	Chukyo metropolitan area PT survey	Kinki metropolitan area PT survey
Survey year	2008	2011	2010
Number of total person trips	587,434	200,916	808,794
Target data	Weekday shopping behaviors Routes not including stopover behaviors (home → shopping place → home only)
Number of target person trips	61,337	11,498	11,773
Survey items	Age, sex, occupation, transportation means, departure time, travel time, travel distance, and duration of time at the destination

The Tokyo metropolitan area includes Tokyo Metropolis, all of Kanagawa, Chiba, and Saitama prefectures, as well as part of Ibaraki Prefecture. The Chukyo metropolitan area includes Aichi, Gifu, and Mie prefectures. The Kinki metropolitan area includes Osaka, Shiga, Kyoto, Hyogo, Nara, and Wakayama prefectures.

Figure 1. Analysis target areas.

Figure 2. Aggregations of extracted shopping behaviors when undertaking shopping trips according to metropolitan area in terms of: (a) Sex, (b) Age, (c) Occupation, (d) Means of transportation (multiple responses allowed), (e) Departure time, (f) travel time (one-way average), (g) Travel distance (one-way average and straight-line distance calculated by coordinate), and (h) Stay time.

Figure 3. Aspects of undertaken shopping trips according to metropolitan area: (a) Distribution of travel time, (b) Distribution of stay time, and (c) Distribution of stay time/travel time ratio.

Figure 4. Distributions of travel and stay times on a natural log scale in the (a) Tokyo, (b) Chukyo, and (c) Kinki metropolitan areas. Note that transparency is assigned to each point according to data concentration, with darker points indicating a larger quantity of data.

Figure 5. Shopping behavior pattern clusters in the (a) Tokyo, (b) Chukyo, and (c) Kinki metropolitan areas.

Table 2. Number and percentage of cases in each cluster.

Area	Early departure and home surrounding type	Early departure and long trip type	Early departure and long stay type	Late departure and home surrounding type
Tokyo	25,484 (42.43%)	7,006 (11.66%)	2,101 (3.50%)	25,472 (42.41%)
Chukyo	5913 (52.30%)	337 (2.98%)	753 (6.66%)	4303 (38.06%)
Kinki	4531 (38.58%)	49 (0.42%)	969 (8.25%)	6196 (52.75%)

The cluster analysis reported in this table excluded the “other/unknown” occupation.

Table 3. Relationship between travel time and stay time.

Regression analysis	Variable	B	β	Significance probability	VIF
Tokyo (R^2 = 0.916)	Travel time (min) (ln)	1.503	0.957	***	1.000
Chukyo (R^2 = 0.933)	Travel time (min) (ln)	1.506	0.966	***	1.000
Kinki (R^2 = 0.922)	Travel time (min) (ln)	1.453	0.960	***	1.000

B is the nonstandard regression coefficient (stay coefficient), β is the standard regression coefficient, VIF is the variance inflation factor, R² is the adjusted coefficient of determination, and *** indicates p < 0.001.

Table 4. Relationship between all variables and stay time.

Regression analysis	Variable	B	β	Significance probability	VIF
Tokyo (R^2 = 0.948)	Male (D)	0.020	0.002	*	1.671
	10s (D)	0.005	0.000	-	5.101
	20s–30s (D)	0.159	0.018	***	1.672
	40s–50s (D)	0.171	0.025	***	2.145
	Employed (D)	0.385	0.051	***	2.404
	Student (D)	0.419	0.021	***	5.215
	Homemaker (D)	0.314	0.054	***	3.561
	Walking (D)	0.282	0.053	***	3.585
	Bicycle (D)	0.354	0.052	***	2.364
	Motorcycle (D)	0.329	0.011	***	1.061
	Car (D)	0.419	0.059	***	2.395
	Bus/Taxi (D)	0.152	0.010	***	1.189
	12 am–6 am (D)	1.594	0.013	***	1.012
	6 am–12 pm (D)	1.414	0.232	***	5.312
	12 pm–6 pm (D)	1.283	0.238	***	6.020
	Travel time (min) (ln)	0.729	0.464	***	9.847
Chukyo (R^2 = 0.961)	Male (D)	0.008	0.001	-	1.658
	10s (D)	−0.166	−0.005	-	4.785
	20s–30s (D)	0.099	0.011	***	1.576
	40s–50s (D)	0.098	0.015	***	2.101
	Employed (D)	0.236	0.030	***	2.315
	Student (D)	0.468	0.016	***	4.840
	Homemaker (D)	0.216	0.039	***	3.689
	Walking (D)	0.238	0.033	***	2.139
	Bicycle (D)	0.415	0.051	***	2.041
	Motorcycle (D)	0.562	0.014	***	1.044
	Car (D)	0.601	0.126	***	5.813
	Bus/Taxi (D)	0.223	0.010	***	1.130
	12 am–6 am (D)	1.082	0.009	***	1.016
	6 am–12 pm (D)	1.455	0.275	***	8.141
	12 pm–18 pm (D)	1.331	0.228	***	6.613
	Travel time (min) (ln)	0.677	0.434	***	11.779
Kinki (R^2 = 0.954)	Male (D)	0.110	0.017	***	1.884
	10s (D)	0.372	0.027	***	8.874
	20s–30s (D)	0.319	0.039	***	2.050
	40s–50s (D)	0.288	0.049	***	2.730
	Employed (D)	0.405	0.070	***	3.434
	Student (D)	0.232	0.016	**	8.936
	Homemaker (D)	0.364	0.052	***	2.596
	Walking (D)	0.299	0.062	***	2.683
	Bicycle (D)	0.272	0.048	***	1.994
	Motorcycle (D)	0.258	0.015	***	1.087
	Car (D)	0.349	0.063	***	2.400
	Bus/Taxi (D)	0.171	0.009	***	1.077
	12 am–6 am (D)	1.404	0.012	***	1.008
	6 am–12 pm (D)	1.071	0.133	***	2.466
	12 pm–18 pm (D)	0.934	0.200	***	4.532
	Travel time (min) (ln)	0.745	0.493	***	8.641

Note: In this table, B is the nonstandard regression coefficient (stay coefficient), β is the standard regression coefficient, VIF is the variance inflation factor, R² is the adjusted coefficient of determination, *** indicates p < 0.001, ** indicates p < 0.01, * indicates p < 0.05, and (D) is a dummy variable.

Table 5. Relationship between travel time and stay time according to shopping behavior pattern.

Regression analysis		Early departure and home surrounding type	Early departure and long trip type	Early departure and long stay type	Late departure and home surrounding type
Tokyo	B	1.571	1.255	2.040	1.536
	R^2	0.927	0.964	0.919	0.916
Chukyo	B	1.507	1.133	1.729	1.521
	R^2	0.935	0.967	0.963	0.932
Kinki	B	1.436	0.937	1.431	1.498
	R^2	0.924	0.963	0.951	0.915

In this table, B is the nonstandard regression coefficient (stay coefficient), R² is the adjusted coefficient of determination, and the bold numbers indicate the maximum values for each trip type in each metropolitan area.

Data availability statement

The data that support the findings of this study are available from the CSIS of the University of Tokyo, but restrictions apply to the availability of these data, which were used under license for the current study and are not publicly available. Data are however available from the author upon reasonable request and with permission of CSIS.