Zenith Model Recalibration and Validation Version 3.0.1 Paper 3 Household Segmentation Model May 2014 Public Transport Victoria
Page Intentionally Left Blank
Paper 3 Household Segmentation Model Draft Report Project No. ZML-VIC-Year4 COPYRIGHT: The concepts and information contained in this document are the property of Veitch Lister Consulting Pty Ltd. Use or copying of this document in whole or in part without the written permission of Veitch Lister Consulting constitutes an infringement of copyright. LIMITATION: This report has been prepared on behalf of and for the exclusive use of Veitch Lister Consulting Pty Ltd s Client, and is subject to and issued in connection with the provisions of the agreement between Veitch Lister Consulting and its Client. Veitch Lister Consulting accepts no liability or responsibility whatsoever for or in respect of any use of or reliance upon this report by any third party. Date Revision Prepared By Checked By Approved By Description 16/03/2014 A MP JC TV Draft Report 02/05/2014 B MP JC TV Draft Report i
Executive Summary The Zenith Model of Victoria is one of a family of models developed by Veitch Lister Consulting (VLC) for transport planning in Australian cities and regions. This document is one in a series of working papers that collectively describe the calibration and validation of the Zenith Model of Victoria. In particular, this document describes the Household Segmentation Model. The aim of the Household Segmentation Model is to segment households according to their level of particular attributes (e.g. level of car ownership), given as input an average (zonal) value for each attribute. For example, if a zone has an average car ownership of 1.3 vehicles per household, the Household Segmentation Model calculates that there are likely to be on average 15.5% households with no cars, 48.5% one car households, 28.3% 2 car households and the remaining 7.7% are households with three or more cars. Separate Household Segmentation Model relationships have been estimated for each of the following household variables: Household size; Number of white collar workers; Number of blue collar workers; Number of dependants aged 0 to 17 years; Number of dependants aged 18 to 64 years; Number of dependants aged 65 years and older; and Number of cars owned. The Household Segmentation Model has been estimated using data obtained from the Australian Bureau of Statistics (ABS) Census (2006 and 2011). Data was obtained for each SA1 / CCD in Victoria. For example, in the case of the variable number of cars owned, this data included: average number of cars per household in each SA1; number of 0 car households in each SA1; number of 1 car households in each SA1; number of 2 car households in each SA1; and number of 3+ car households in each SA1. Using this data for each of the SA1s / CCDs in Victoria, a series of mathematical relationships (of Logit form) have been estimated. For example, in the case of car ownership, the following curves were fit to the ABS data. ii
Figure 0-1 - Segmentation of Households by Number of Cars Owned In summary, the above Figure can be interpreted as follows: For each SA1: the average number of cars owned by households resident in that area was calculated and then used as the x-axis value; and The fraction of households resident in that area owning 0 cars, 1 car or less, and 2 cars or less was calculated and then plotted along the y-axis, resulting in 3 data points (blue, yellow and green, respectively). The black lines represent the fitted model for each level; The shaded regions represent the expected number of households owning 0 cars, 1 car, 2 cars, and 3+ cars. Using these relationships, it is possible to estimate, for any given average car ownership level, the fraction of households owning 0 cars, 1 car, 2 cars or 3+ cars. Similar relationships were found for the other variables and the resulting relationships developed and documented in this paper have been implemented into the Zenith Model of Victoria. iii
Contents Executive Summary... ii Contents... iv List of Figures... v List of Tables... vi 1 Introduction... 1 2 Data Sources... 2 3 Data Analysis... 3 4 Parameter Estimates... 8 iv
List of Figures Figure 0-1 - Segmentation of Households by Number of Cars Owned... iii Figure 3-1 - Segmentation of Households by Number of Residents... 4 Figure 3-2 - Segmentation of Households by Number of Workers (White Collar)... 4 Figure 3-3 - Segmentation of Households by Number of Workers (Blue Collar)... 5 Figure 3-4 - Segmentation of Households by Number of Dependants (Aged 0-17)... 5 Figure 3-5 - Segmentation of Households by Number of Dependants (Aged 18-64)... 6 Figure 3-6 - Segmentation of Households by Number of Dependants (Aged 65+)... 6 Figure 3-7 - Segmentation of Households by Number of Cars Owned... 7 v
List of Tables Table 1 - Re-estimated Household Segmentation Parameters... 8 vi
Page Intentionally Left Blank vii
1 Introduction The Zenith Model of Victoria is one of a family of models developed by Veitch Lister Consulting (VLC) for transport planning in Australian cities and regions. This document is one in a series of working papers that collectively describe the calibration and validation of the Zenith Model of Victoria. The subject of this working paper is the Household Segmentation Model developed for the Zenith Model of Victoria, including data analysis and estimated model parameters. This working paper does not include a description of the role and methodology of the Household Segmentation Model. This information can be found in the Zenith framework document Zenith Framework Household Segmentation. It is recommended that the framework document be read prior to reading this working paper. 1
2 Data Sources The Household Segmentation Model for the Zenith Model of Victoria has been estimated using a mixture of data from the 2006 ABS Census and the 2011 ABS Census. The data source used for each variable is as follows: Household size ABS Census 2011 Number of white collar workers ABS Census 2006 Number of blue collar workers ABS Census 2006 Number of dependants aged 0-17 ABS Census 2006 Number of dependants aged 18-64 ABS Census 2006 Number of dependants aged 65+ ABS Census 2006 Number of cars owned ABS Census 2011 It was necessary to use 2006 data for some variables (rather than 2011 data) because of difficulties in obtaining data from the Australian Bureau of Statistics (ABS) within the timeframes of this update. A request has been submitted to the ABS which will allow the full update of all variables using 2011 data, and a revised version of this working paper will be produced shortly after that data is received. 2
3 Data Analysis The data received from the ABS is plotted in Figure 3-1 through Figure 3-7 below. Referring to Figure 3-1 (household size) as an example: For each Victorian statistical area (SA1 for 2011 and CCD for 2006): the average household size for households resident in that area was calculated and then used as the x-axis value; and The fraction of households resident in that area having 1 person, 2 persons or less, 3 persons or less, 4 persons or less, and 5 persons or less was calculated and then plotted along the y-axis, resulting in 5 data points (blue, yellow, green, red and purple). The black lines represent the fitted model for each level; and The shaded regions represent the number of households having 1 person, 2 persons, 3 persons, 4 persons, 5 persons and 6+ persons. As an example of how this model would be applied, if the average household size for a given travel zone was 2.4 persons per household, then the model would predict the following distribution of household sizes: 1 person 27% of households; 2 persons 36% of households; 3 persons 16% of households; 4 persons 13% of households; 5 persons 6% of households; and 6+ persons 2% of households. 3
Figure 3-1 - Segmentation of Households by Number of Residents Figure 3-2 - Segmentation of Households by Number of Workers (White Collar) 4
Figure 3-3 - Segmentation of Households by Number of Workers (Blue Collar) Figure 3-4 - Segmentation of Households by Number of Dependants (Aged 0-17) 5
Figure 3-5 - Segmentation of Households by Number of Dependants (Aged 18-64) Figure 3-6 - Segmentation of Households by Number of Dependants (Aged 65+) 6
Figure 3-7 - Segmentation of Households by Number of Cars Owned 7
4 Parameter Estimates The model parameters estimated for the Household Segmentation Model are presented in Table 1 below. T values are provided for each curve. Trip Type Model Coefficients Household Size Workers (White Collar) Workers (Blue Collar) Dependants (Aged 0-17) Dependants (Aged 18-64) Dependants (Aged 65+) Cars Table 1 - Re-estimated Household Segmentation Parameters 8