Figure 2.

Comparison of network buffer methods to evaluation of local neighbourhood land use composition. Four examples below compare the buffered line-based network buffer ('buffered line method') to the polygon-based network buffer ('polygon method'). (A) The use of the polygon method adds a substantial amount of area to the local neighbourhood that is not actually accessible by an individual. (B) Using the buffered line method, it is evident that this individual would experienced his/her neighbourhood as being overwhelmingly residential but the polygon method would understate this experience by adding two large areas that cannot be meaningfully interacted with. In particular the large region to north of this individual appears to be completely concealed from the nearest roads by the houses that line the streets – its inclusion in the polygon method will greatly overestimate the presence of green space relative to a model focused on where the individual can actually cover walking. (C) The percentages of both institutional and 'other' land are greater when using the polygon method, thereby decreasing the relative weight given to land more practically accessible. (D) The polygon method may overstate the industrial land's importance within this individual's local neighbourhood, in terms of the influence on walking.