Changing Spatial Patterns in the Journey-to-Work : a comparison of the 1966 and 1971 Census Data in London

Abstract

The 1966 and 1971 census journey-to-work data for London has been analysed by traffic zone of origin and destination (residence and workplace), using LTS and GLTS traffic zones respectively. The 10 per cent sample, or some 400,000 journeys, is thus analysed by some 1000 zones for each of which coordinate references or geo-codes of their centroids are available. Direct distances for the journey length between the centroids, and for the distance of the centroid from the city centre can thus be calculated, using for the city centre a point near Piccadilly Circus which is the centre of population. These distances can be chopped into 2 km bands, and the following analyses performed: (a) median trip lengths can be obtained, by distance of the trip end from the city centre, in total and by mode; (b) numbers of workplaces and residences, by distance from the centre, in total and by mode; (c) modal split, by distance of the trip end from the city centre; (d) the workplace-based trip length distributions are calibrated by gamma distributions (but not the residence-based). The results show an increase in median trip lengths of 2-3 per cent per annum overall, over the 5 years, but the detailed switching between modes as car ownership increases shows that, by mode, the trip length distributions are much more constant, only rail and car trips from workplaces showing any substantial increase (of 2 per cent per annum). Goodwin's theory of utility implies a particular spread factor for the gamma calibrations of trip length. The observed distributions do not conform to this. These suggest that for bus trips the taper of generalised cost with distance is important, and more crucially for city structure, that for rail and car travel the trade-off between rent and transport costs is important, giving radial rather than orbital travel.