Date of Award


Document Type

Union College Only

Degree Name

Bachelor of Arts



First Advisor

Yufei Ren




traffic, competition, data, high, location


This paper studies the spatial distribution of coffee shops in a specific city center - San Francisco. Classical spatial competition theory maintains a wide moat from empirical data. My goal is to examine, adjust or add to this theory in the light of hard information that is available for a particular scenario. Lacking access to business revenue data, I focus on a different, increasingly relevant form of currency – online reviewer traffic. I take reviewer traffic (I) primarily as a proxy for business traffic, and thus for revenue, and (II) secondarily as something that is desirable in itself as a form of market exposure. Using over 700 observations of manually collected data, I find that location choice relative to competition does matter in a predictable and meaningful way. Since the initial question of location choice presents a tradeoff – on the one hand, high demand areas offer strong currents of potential customers; on the other hand, high demand areas also promise strong competition from other businesses – the overall effect of location is theoretically indeterminate. Working carefully with data from an unusual perspective, I develop algorithms to measure key atmospheric competition characteristics, and succeed in predicting the effects of location on reviewer traffic with robust statistical significance. In particular, high atmospheric traffic is beneficial, while high atmospheric quality is detrimental – both of which can be measured anywhere on the map.