The Influence of Built Environment and Socio-Economic Factors on Commuting Energy Demand: A Path Analysis-Based Approach

Boukarta, Soufiane; Berezowska-Azzag, Ewa

doi:10.2478/quageo-2022-0039

Figures & Tables

Study area and the neighbourhoods of the city according to their period of creation.Source: own composition based on Bouzidi et Boukhari 2013.

Conceptual model explaining the need for commuting.

The hypothesised model.In red: positive causal relation; in black: negative causal relationSource: own study.

Comparison between the survey and census data.LPG – liquefied petroleum gasSource: own study.

Employment and housing locations.green: workplace, yellow: housing, white: centre of the citySource: authors’ compilation based on Google Earth (2015).

Direct effects of driving factors explaining the energy consumption generated by commuting. R2 = 0.61.Source: own compilation.

Effect of SE and BE driving factors per unit of measurement on energy consumption of commuting to work.BE – built environment, SE – socio-economicSource: own compilation.

Direct and indirect effects of the driving factors of energy consumption generated by commuting_

	Effect	Number of cars	Home-to-work distance	Round trip frequency	Profession	Built density	Energy consumption (kWh person⁻¹ × year⁻¹)
Number of floors	Direct	0.255	0.000	0.000	0.000	0.000	0.000
	Indirect	0.000	−0.030	0.009	0.000	0.000	0.097
	Total	0.255	−0.030	0.009	0.000	0.000	0.097
Housing type	Direct	0.295	0.000	−0.139	0.227	0.664	0.000
	Indirect	0.000	−0.035	0.010	0.000	0.000	0.016
	Total	0.295	−0.035	−0.129	0.227	0.664	0.016
Income	Direct	0.425	0.000	0.000	0.000	0.000	0.000
	Indirect	0.000	−0.050	0.014	0.000	0.000	0.161
	Total	0.425	−0.050	0.014	0.000	0.000	0.161
Occupancy rate per housing	Direct	0.000	0.130	0.000	0.000	0.000	0.000
	Indirect	0.000	0.000	−0.037	0.000	0.000	0.076
	Total	0.000	0.130	−0.037	0.000	0.000	0.076
Bus rotation	Direct	0.000	0.427	0.000	0.000	0.000	0.000
	Indirect	0.000	0.000	−0.122	0.000	0.000	0.249
	Total	0.000	0.427	−0.122	0.000	0.000	0.249
Distance to centre	Direct	0.000	0.330	0.000	0.000	−0.331	0.000
	Indirect	0.000	0.000	−0.094	0.000	0.000	0.241
	Total	0.000	0.330	−0.094	0.000	−0.331	0.241
Number of cars	Direct	0.000	−0.117	0.000	0.000	0.000	0.448
	Indirect	0.000	0.000	0.033	0.000	0.000	−0.068
	Total	0.000	−0.117	0.033	0.000	0.000	0.379
Education level	Direct	0.000	0.000	0.000	0.000	0.111	0.000
	Indirect	0.000	0.000	0.000	0.000	0.000	−0.016
	Total	0.000	0.000	0.000	0.000	0.111	−0.016
Bus frequency	Direct	0.000	0.000	−0.105	−0.155	0.000	0.000
	Indirect	0.000	0.000	0.000	0.000	0.000	−0.054
	Total	0.000	0.000	−0.105	−0.155	0.000	−0.054
Distance to national road	Direct	0.000	0.000	0.000	0.000	0.252	0.000
	Indirect	0.000	0.000	0.000	0.000	0.000	−0.037
	Total	0.000	0.000	0.000	0.000	0.252	−0.037
Home-to-work distance	Direct	0.000	0.000	−0.285	0.000	0.000	0.661
	Indirect	0.000	0.000	0.000	0.000	0.000	−0.077
	Total	0.000	0.000	−0.285	0.000	0.000	0.584
Respondent age	Direct	0.000	0.000	0.000	0.000	0.000	−0.119
	Indirect	0.000	0.000	0.000	0.000	0.000	0.000
	Total	0.000	0.000	0.000	0.000	0.000	−0.119
Round–trip frequency	Direct	0.000	0.000	0.000	0.000	0.000	0.271
	Indirect	0.000	0.000	0.000	0.000	0.000	0.000
	Total	0.000	0.000	0.000	0.000	0.000	0.271
Profession	Direct	0.000	0.000	0.000	0.000	0.000	0.165
	Indirect	0.000	0.000	0.000	0.000	0.000	0.000
	Total	0.000	0.000	0.000	0.000	0.000	0.165
Built density	Direct	0.000	0.000	0.000	0.000	0.000	−0.145
	Indirect	0.000	0.000	0.000	0.000	0.000	0.000
	Total	0.000	0.000	0.000	0.000	0.000	−0.145

Direct effects between driving factors_

			Effect	P
Number of cars	< – -	Income	0.425	^***
Number of cars	< – -	Number of floors	0.255	0.019^**
Number of cars	< – -	Housing type	0.295	0.004^**
Home-to-work distance	< – -	Distance to centre	0.330	^***
Home-to-work distance	< – -	Number of bus rotation	0.427	^***
Home-to-work distance	< – -	Occupancy rate per housing	0.130	0.070^*
Home-to-work distance	< – -	Number of cars	−0.117	0.071^*
Built density	< – -	Distance to national road	0.252	0.014^**
Profession	< – -	Bus frequency	−0.155	0.032^**
Built density	< – -	Distance to centre	−0.331	0.002^**
Built density	< – -	Education level	0.111	0.061^*
Profession	< – -	Housing type	0.227	0.002^**
Round trip frequency	< – -	Housing type	−0.139	0.057^*
Built density	< – -	Housing type	0.664	^***
Round trip frequency	< – -	Bus frequency	−0.105	0.149
Round trip frequency	< – -	Home-to-work distance	−0.285	^***
Energy consumption (kWh × person⁻¹ × year⁻¹)	< – -	Home-to-work distance	0.661	^***
Energy consumption (kWh × person⁻¹ × year⁻¹)	< – -	Profession	0.165	^***
Energy consumption (kWh × person⁻¹ × year⁻¹)	< – -	Round trip frequency	0.271	^***
Energy consumption (kWh × person⁻¹ × year⁻¹)	< – -	Number of cars	0.448	^***
Energy consumption (kWh × person⁻¹ × year⁻¹)	< – -	Built density	−0.145	0.007^**
Energy consumption (kWh × person⁻¹ × year⁻¹)	< – -	Respondent age	−0.119	0.032^**

Descriptive statistics_

Variables	N	Variable type	Minimum	Maximum	Average	SD
Accessibility
Outward journey time (min)	175	Continuous	2	200	20.59	18.26
Home–work distance (m)	162	Continuous	50	18,000	2,206.91	1,966.96
Number of bus rotations	150	Continuous	0	3	1.20	0.556
Density
Plot ratio^*	139	Continuous	0.17	1.00	0.59	0.32
Built density^*	139	Continuous	0.60	3.20	2.02	0.74
Design
Distance to centre (m)^*	148	Continuous	17.59	2,659.18	1,219.54	696.57
Distance from national road (m)^*	151	Continuous	25.62	2,569.23	1,163.83	629.94
Average number of floors. (n)^*	139	Continuous	2.00	6.00	3.79	1.08
Block's area (m²)^*	139	Continuous	770	36.061	7,398.83	9,150.32
Housing type (1: collective, 2: individual)	184	Nominal	1	2	1.54	0.50
Distance to public transport
Distance to public transport (housing zone)(0: <300, 4: >1 km)	173	Ordinal	1.00	4.00	2.3237	0.98
Distance to public transport (work zone)(0: <300, 4: >1 km)	172	Ordinal	0.00	4.00	1.6919	0.97
Bus frequency	184	Continuous	0.00	5.00	2.4620	1.56
Diversity
Mixed use index (from 5 to 40)^*	175	Continuous	12	36	24.23	5.38
Households’ SE characteristics
Household's average age^*	42	Continuous	16.33	43.80	27.1681	8.44
Respondent age	120	Continuous	27	70	43.95	10.82
Round-trip frequency	172	Continuous	1	4	1.66	0.51
Household's education level	46	Continuous	2.00	5.00	3.5230	0.77
Respondent's education level	174	Ordinal	0	4	3.26	1.06
Number of cars owned	184	Continuous	0	2	0.60	0.57
Profession (1: public, 2: liberal)	181	Nominal	1	3	1.19	0.52
Income (from 15,000 to + 60,000 Da)	181	Ordinal	1	4	2.15	0.95
Occupancy rate per housing.	139	Continuous	2	12	5.34	1.87
Modal share
Public transport (TC (1: TC, 0: other)	184	Nominal	0.00	1.00	0.31	0.46
Car (1: Voiture, 0: other)	184	Nominal	0.00	1.00	0.26	0.44
Walking (1: MAP, 0: other)	184	Nominal	0.00	1.00	0.42	0.49

Fuel conversion to kWh per km per person_

	Means of commuting
	Car			Bus
Fuel type	Diesel	Petrol	LPG	Diesel
Consumption (L × km⁻¹)	0.063^*	0.075^*	0.075^*	0.2^**
Rate of occupation per vehicle	1.27^*	1.27^*	1.27^*	28^**
Density (Tonne/m³)	0.825^****	0.735^*****	0.55	0.825
Conversion factor 1 (T fuel – >Toe)^***	1,015	1,054	1,084	1,015
Conversion factor 2 (Toe – >KWh)	11,630^***
Consumption KWh × km⁻¹	0.61	0.68	0.52	1.95
Consumption KWh per person per Km per one-way trip	0.48	0.53	0.41	0.07
Consumption KWh per person per Km per year (225 day)	108	119.25	92.25	15.75

Fit indices of the energy consumption of commuting model_

	df	c²	Probability level	RMSEA	NFI	CFI
Commuting energy consumption model	53	47.785	0.677	0.000	0.948	1.000

Description of the papers chosen for the literature review_

Authors	Country	Period	Data sources	Study scale	Sensitivity analysis method	Sample size	Mobility type			drivers		explanatory power of the modal
Authors	Country	Period	Data sources	Study scale	Sensitivity analysis method	Sample size	C	S	L	SE	BE	explanatory power of the modal
Van Acker and Witlox (2010)	Belgium	2000–2001	Survey on behaviour of travellers in Ghent on people aged 18 and over.	City	SEM	2,500 households	×	×	×	×	×	R² = 20.1%
Breheny (1995)	UK – Wales	1961–1991	Aggregated data from Ecotec project (1993).	National	Interpolation from data from Ecotec project (1993).	Ecotec project sample (1993)	×	×	×		×	–
Brownstone and Golob (2008)	California	2001	National Household Transportation Survey. Aggregate data.	National	SEM	2,079 households	×	×	×	×	×	R² = 0.37 and 0.42
Calabrese et al. (2012)	Massachusetts, USA	2011	Deducted by detecting signal of mobile phones carried out by AirSag.	Metropolitan area	Multiple linear regression	1,101 households	×	×	×	×	×	R² = 49.40% and 56.48%
Newman et al. (1989)	32 cities of different countries	1980	Collection of fuel consumption data and calculation of density excluding rural areas. Urban planning agency of different countries. Aggregate data.	City	Bivariate correlation analysis	32 cities	×	×	×	×	×	/
Cervero and Murakami (2010)	USA	2003	Data collected from Highway Statistics. Department of Commerce.	National	SEM	370 urban areas	×	×	×		×	CFI (>0: 900) 0.969NFI (>0: 950) 0.961NNFI (>0: 900) 0.942
Cervero and Radisch (1995)	USA	1990–1991	Bay Area Travel questionnaire survey.	Neighbourhood	Binary logistic regression	2 Neighbourhoods: 620 households for non commuting. And 840 households for commuting	×	×	×	×	×	Pseudo R² = 0.29,Predicted cases = 88.6%.
Chen et al. (2007)	NY, USA	1997/1998	Household survey	Metropolitan area	SEM	2,089 trips	×			×	×	R² = 0.45 and 0.58
Dargay (2004)	UK	1970–1995	Surveys of family spending.	National	Semi-logistic regression	256 pseudo panels	×	×	×	×	×	R² = 0.989
Dieleman et al. (2002)	Netherlands	1996	National Mobility Survey in the Netherlands	National	Multinomial logistic regression	70,000 households	×	×	×	×	×	R² = 0.31
Ding et al. (2017)	Baltimore USA	2001	Household survey	Metropolitan area	SEM	3,519 households	×			×	×	/
Feng et al. (2013)	China and Netherlands	2008	Household survey on mobility in both countries.	City	Multiple linear regression	2,989 respondents for 10 districts in China and 1,322 respondents for Randstad.	×	×	×	×	×	China:R² = 0.115RandstadR² = 0.124
Handy et al. (2005)	California (US)	2003	E-mail questionnaire carried out on eight neighbourhoods.	District in metropolitan area	Linear regression	1,466 respondents				×	×	R² = 0.16R² adjusted = 154
Holden and Norland (2005)	Oslo, Borway	2003	Questionnaire distributed by mail.	Regional	linear regression	650 for daily trips, 778 for leisure travel, <100 respondents per zone (eight zones selected for the study).	×		×	×	×	R² = 0.231 for commuting
Karathodorou et al. (2010)	42 countries	1995	Millennium Cities Database for Sustainable Transport (1999) for 100 countries. And car occupancy from Mobility in Cities database (2006).	Cities	Linear regression	84 cities	×		×	×	×	R² = 0.61
Khan et al. (2014)	Seatle, USA	2006	Questionnaires/Puget Sound Regional Council	Metropolitan area	Regression modelling	10,510 respondents of 4,741 households.	×		×	×	×	/
Kitamura et al. (1997)	San Francisco, USA	1994	Questionnaire, And land use information is obtained from the Metropolitan Transportation Commission.	Neighbourhood	Multiple linear regression	5 Neighbourhoods, 640 respondents,	×		×	×	×	R² = 0.2125
Limtanakool et al. (2006)	Netherlands	1996	National Mobility Survey conducted by telephone interview and questionnaire	Regional	Binary logistic regression	Commuting: 2,326Shopping: 932Leisure: 3,072	×	×	×	×	×
Ma et al. (2014)	China	2007	Questionnaires	Neighbourhoods	Logistic regression	60 households, 699 trips of 10 neighbourhoods.	×	×	×	×	×	Pseudo R² = 0.16
Manaugh et al. (2009)	Montréal, Canada	2003	Origin-destination survey,	Neighbourhoods	Linear regression	17,000 trips	×			×	×	SE: R² = 0.06.SE+BE modal:R² = 0.40
Marique (2013)	Belgium	2001	2001 SE survey	National	Multiple linear regression	966.247 respondents.	×			×	×	R² = 0.457
Næss (2010)	Hangzhou, China	2005	Qualitative interview and questionnaire in 40 urban areas.	Urban zone	Multiple linear regression	28 interviews3,150 questionnaire respondents.	×			×	×	R² = 0.189
Naess (2014)	Hangzhou, China and Copenhague, Danemark	2005	Interview and questionnaire	Regional	Linear regression	1932 et 3150 questionnaire	×			×	×	CopenhagueR² = 0.233HangzhouR² = 0.095
Pan et al. (2009)	Shanghai, China	2001	Questionnaires	Neighbourhood	Multiple logistic regression	1,709 respondents in 4 Neighbourhoods	×		×	×	×	Pseudo R² = 0.2714
Zhang et al. (2014)	Zhongshan, China	2010	Questionnaires	Neighbourhoods	Linear regression	25,618 respondents	×	×	×	×	×	Pseudo R² = 0.2823
Bakour (2016)	Algiers, Algeria	2004	Household survey conducted by an organisation	City	Linear regression	1,200 respondents	×					R² = 0.5 à 0.9

Pearson's bivariate correlation_

	Commuting energy consumption (KWh × person⁻¹ × year⁻¹)
	Correlation	Sig	No.
Home-work distance (m)	0.561^*	0.000	162
Outward journey time (min)	0.035^*	0.661	163
Number of bus rotations	0.247^*	0.004	136
Built density	−0.135^*	0.133	125
Plot ratio	−0.120^*	0.184	125
Housing type (1: collective, 2: individual)	−0.153^*	0.051	164
Distance from national road (m)	0.256^*	0.003	136
Distance to centre (m)	0.280^*	0.001	133
Block's area (m²)	0.224^*	0.012	125
Average number of floors	0.143^*	0.112	125
Mixed use index	−0.049^*	0.545	156
Distance to public transport (housing zone)	−0.152^*	0.059	154
Distance to public transport (work zone)	−0.145^*	0.064	164
Round trip frequency	0.118^*	0.141	156
Profession (1: public, 2: liberal)	0.169^*	0.031	163
Respondent age	−0.265^*	0.005	109
Respondent's education level	0.107^*	0.183	156
Household's education level	0.102^*	0.541	138
Income	0.136^*	0.082	163
Household's average age	−0.307^*	0.064	137
Number of cars owned	0.379^*	0.000	163
Occupancy rate per housing	−0.073^*	0.418	126

Descriptive statistics of the energy consumption generated by commuting to work_

	Minimum	Maximum	Average	SD
Home-to-work distance	50	18,000	2,206.91	1,966.96
Daily consumption (kWh × person⁻¹)	0.00	3.60	0.4853	0.788
Annual consumption (kWh × person⁻¹ × an)	0.00	810.00	109.1854	177.31

The Influence of Built Environment and Socio-Economic Factors on Commuting Energy Demand: A Path Analysis-Based Approach

Figures & Tables

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5

Fig. 6

Fig. 7

Direct and indirect effects of the driving factors of energy consumption generated by commuting_

Direct effects between driving factors_

Descriptive statistics_

Fuel conversion to kWh per km per person_

Fit indices of the energy consumption of commuting model_

Description of the papers chosen for the literature review_

Pearson's bivariate correlation_

Descriptive statistics of the energy consumption generated by commuting to work_

Paradigm

My account