 |
 |
There are different aims at stake; one is explanatory power and theory building the other one is prediction and pragmatic uses, the later with a more technological flavour. Hopefully the former will lead to the latter but that may take a lot of time, but these two strands are not always in synch. There is a difference between models where independent and dependant variables are kept separated and a model where output variable are used as the input variable. When you mix the two, it becomes a kind of sophisticated trend extrapolations device and surely not quite an explanatory model but a pragmatic one, most useful as such if nothing else is available. You can have very convincing animation of water flow that match actual flow but it does not mean that you have an explanatory model - to predict is not the same than to explain (quoting Rene Thom). From Bill's explanation; the Nantes space syntax multivariable vehicular traffic model includes a managed variable: road width, which play an important role too, in that sense they will be no difficulty to add a dummy variable as part of the managed side of the equation like road pricing and see what happen. Andrew - if you get us the data, we are ready to have a go - we would need useable road width data, but that could be easily retrieved from GIS, plus count inside and outside the Congestion Charging Zone, before and after. It would require much more work for bus lanes, traffic light phasing, traffic calming. Bus lanes are about car useable road width reduction so lets leave that for much later. Rome was not build in one day. So lets be fair. "There's little new in the Boston conclusions in the document that was linked to, earlier. I'd say that it seems to be somewhat behind the times IMHO. The transport planning profession has known about the linkages between land use mixture, density and travel demand for over 40 years (not that it's been reflected in policy and action, but nevertheless the knowledge has been there), and there's been lots of very good work done in the last 4 decades on describing those linkages." Absolutely right, there is nothing new there after 40 years of research. What I was suggesting was that the known linkage (this is again different from explaining how the coupling is occurring apart from a circular gravity model) between density, land use mix, travel demand process changes seems to be better explained and probably quantified from a spatial accessibility process change point of view. Urban form dynamic is a conflict of attractors, and spatial accessibility seems to be an extremely powerful one. I agree there is understanding - There is a difference between understanding the role of spatial accessibility if that is what you mean by linkage (and what of it) and having a parameter in a model that includes its quantified effect. "As traffic (or more usefully, general transport) models use origin/destination matrices, I can't see how one would implement Alan Penn's suggestion of: "using space syntax measures ... in the matrix estimation phase of the construction of traffic models", but am willing to be enlightened. I'm often agnostic as to the usefulness of models based on such O/D matrices, but they do have lots of advantages in terms of being able to inform policy decisions on land use and transport." I would agree with you on the usefulness of such model because of all the general 4 steps that traffic model require the origin/destination matrix (the second step which is before mode assignment if I recall correctly) is probably the most difficult to justify objectively. Very recent and long conversations with our friends transport engineers/planners at JMP and Steer Davies Gleaves led to the admission that this is more a well established and accepted convention protocol than anything they could justify the detail of. And I will go with that. Along that line, model will not change if for example TfL is continuing to only accept to support one sort of model. So this is quite an interesting social issue, the institutionalisation of knowledge blocking innovation, not new either (see Kuhn). And so there is little doubt that it would be very difficult to move the transportation-modelling establishment - this is like trying to manoeuvre a super tanker with your little finger, as you said 40 years of worldwide research, hundred of thousands practitioners, billions worth of research and professionals fees. There is another aspect of modelling that could be queried when comparing model - if we have competing models, one with few variables and one with many more variable with similar explanatory power, which one should you chose? But yes then just for the fair play, we would like to sit down and look at how this is done in the detail and work out with you that probably we have a better alternative that may be then change the usefulness of such models. _______________________________ Alain Chiaradia Associate Director SPACE SYNTAX _______________________________ -----Original Message----- From: Andrew Smith [mailto:[log in to unmask]] Sent: 07 May 2004 19:56 To: [log in to unmask] Subject: Re: What streets to include in axman If you'll permit me stepping into the discussion of space syntax and transport modelling, as a transport modeller / planner with a mere inkling of space syntax. In essence, transport models are *pivot* models. A transport model's utility is in reliably forecasting the *change* from a current situation that a policy or scheme would have, rather than in describing the status quo. So a model that explains 80% of the traffic flows in Central London (before congestion charging), based on an axial map, will predict no changes after such road pricing is introduced, as the axial map does not change. Similarly with bus lanes, traffic light phasing, traffic calming: all the standard tools of urban traffic management. But flows have changed significantly ... and a useful transport model would predict those changes reliably. > If we combine the Nantes results ... > we won't be has puzzled than the > transport engineer/planner seems to be in > their conclusions. There's little new in the Boston conclusions in the document that was linked to, earlier. I'd say that it seems to be somewhat behind the times IMHO. The transport planning profession has known about the linkages between land use mixture, density and travel demand for over 40 years (not that it's been reflected in policy and action, but nevertheless the knowledge has been there), and there's been lots of very good work done in the last 4 decades on describing those linkages. Total mobility is driven by relative and absolute accessibility of individual modes, of course, and so a useful network model needs to have coded within it which modes can use which links. As traffic (or more usefully, general transport) models use origin/destination matrices, I can't see how one would implement Alan Penn's suggestion of: "using space syntax measures ... in the matrix estimation phase of the construction of traffic models", but am willing to be enlightened. I'm often agnostic as to the usefulness of models based on such O/D matrices, but they do have lots of advantages in terms of being able to inform policy decisions on land use and transport. Andrew Smith