The theory of geographical markets is based on the notion that economic activity is not evenly spread and regional inequalities have an impact on the decisions of economic agents. Retail gasoline markets are almost perfect examples of geographical markets. The gasoline sold by the stations is a fairly homogeneous good and hence, stations compete for customers locally only. In this study we analyze how market structure, geographical attributes and in particular, demand will influence prices. Our work is a natural follow-up to the study of Csorba et al. (2009) who showed that Hungarian retail gasoline prices are affected by conditions of market competition. Compared to the related literature there are two novelties in this study. First, the dataset allows us combining station specific and local attributes to explain prices. With some chains applying different strategies compared to smaller firms, this is an important complementary feature compared to the studies relying on purely region/local market specific attributes. Second, we extend the estimation of the demand side and partly use it as instruments. Local GDP and population density are two common proxies for demand. We extend the analysis from households to contain firms' and transit traffic demand. To estimate the elasticity of demand we use quantitative (population, number of firms and tourists) and income-effect (per capita income, ratio of big firms and foreign tourists) variables. This methodology enables us to instrument station density, used in classical price estimation models, (partly) with demand factors. We find that demand influences prices mainly through the number of stations. There is some local price variation across chains, and the composition of station has a small impact as well. It is an interesting result, which, along with highways and country borders, has an effect on price determination.