Stavroulaki and colleagues move Space Syntax from explanatory analysis toward prospective design intelligence. Their key proposition is methodological parsimony: a useful predictive model need not reproduce the full complexity of pedestrian behaviour if a small set of robust spatial predictors can forecast flows early enough to affect planning decisions. Street-network measures, built density and selected spatial variables are combined through LASSO regression to estimate pedestrian counts in existing and unbuilt conditions. The conceptual operation is significant because it distinguishes explanation from prediction while preserving their relation: the variables that help interpret observed movement become candidates for testing future spatial scenarios. The work therefore shifts evidence from retrospective description toward anticipatory design. Its wider bridge is to planning support systems, machine learning and evidence-based urbanism, where the decisive question is how models can remain simple enough for design practice while rigorous enough to expose the consequences of structural spatial choices before they become materially fixed.
URBANISM MEETS ART