I'd like to get an idea for what's involved in creating isochrones for a series of sites, with a reasonable degree of accuracy to the estimates. A general outline would be great, though my scenario would be for sites in London as I guess the problem varies with availability of data.
--thanks for comments so far, very useful (the london at night mashup's great :D), sorry the q was too vague. In this case I'm looking at travel time modelling for specific sites using transport network vectors and nodes. Data currently available are the OS Master Map Integrated Transport Network (ITN), National Public Transport Data Repository (NPTDR) which contains a snapshot of public transport timetables in "ATCO-CIF format (including its accepted variants) or in TransXChange TXC v2.1 format" (inc. bus stops), also most of the OS datasets. Tools to hand are MapInfo 10.5, Manifold 8, SQL server 08, an open source stack (postgres/pgis/qgis etc.). Trying to get a handle on the scale and complexity of the task, possible approaches and requirements. It looks big, complex and demanding, but on the other hand must be a common application of GIS technologies.