On the Spatial Point Pattern on Customer Affect Retail Fuel Stations Businesses

Retail fuel stations need to plan business development to increase profitability despite high market competition. The patterns of people's habits and fuel supply dynamics have caused income differences in various fuel retail stations. It is difficult for retail fuel stations to increase the sales of fuel oil because of its reliability. This study aims to investigate the pattern of influence of the distribution of consumer locations on the existence of fuel stations, with the expectation used to maximize land productivity at fuel stations. The main contribution of this research is the modeling of spatial point pattern in examining factors affecting customer demography of the facility variability in the closest fuel stations for each district. This can be obtained by applying the stochastics processes couple with Generalized Linear Models (GLMs), namely the regression modeling to the non-homogenous Poisson Process (NHPP) data as the observation related to the customer location and competence of the fuel stations as the counting data. Two modeling structure scenarios were performed based on the type of the response data around 18 district, i.e. homogenous and non-homogenous Poisson model. The best model is chosen by selecting the smallest Akaike Information Criteria (AIC) value. It is found that the non-homogenous model is superior model with the leading AIC at 95% better than the one level model. We found that, in the long term, customer visit intensity has significant impact to number of pump capacity and merchant store. Future research should explore the other covariate and customer choice strategy when refuelling.