Two completely unrelated stories: (1) the reintroduction of traffic wardens to a town suffering 'parking chaos' for the last year; (2) lots of street parties organised to celebrate the Diamond Jubilee.
In both situations individual actions generate large externalities. In the first, individuals parking badly create chaos for others. In the second, individuals organise street parties for their neighbours. In both cases, there are private gains (easy parking, the fun of the party) being traded off against private costs (time, money) and social costs and benefits. Economists like thinking about mechanisms that generate collaborative behaviour in these circumstances, without the need for further regulation. Repeated interaction and a lack of anonymity help encourage such collaboration (because behaving well is more likely to generate reciprocal good behaviour now or in the future). Geographical proximity raises the chances of repeated interactions and reduces anonymity. Hence street parties get organised, but for the town as a whole no traffic wardens leads to parking chaos.
Understanding these kind of collaborative behaviours has implications far beyond parties and parking. For example, some researchers looking at the cluster of firms in the third Italy have argued that such considerations play an important role in explaining the collaboration that underpins success. As always, however, it's far harder to move from an explanation as to why such collaboration might occur to the success of individual firms. It's equally possible that more successful firms may simply choose to be more cooperative and that most of the success is explained by other factors.
In short, the role of space in facilitating collaboration tells us something about clusters, parties and parking. Personally, that eclecticism is one of the main reasons why I find spatial economics so interesting.