Authors: Jialin Li*, The Ohio State University, Ningchuan Xiao, The Ohio State University
Topics: Geographic Information Science and Systems, Economic Geography, Spatial Analysis & Modeling
Keywords: gas price, spatial analysis, spatial co-location pattern, density map
Session Type: Paper
Start / End Time: 8:00 AM / 9:40 AM
Room: Astor Ballroom I, Astor, 2nd Floor
Presentation File: No File Uploaded
Gas prices vary significantly among stations across space and through time. In this study, we use data collected by a member-based website to understand the spatial and temporal patterns of gas prices around Columbus, Ohio. This includes hourly prices for the regular gasoline for the past two years, and at each hour the 15 stations with the highest prices and 15 with lowest are stored. The stations with highest and with lowest prices are treated as two categories and we explore how they co-locate with each other. First, kernel density of the gas stations is estimated for high and low prices, respectively. In addition to mapping the locations with high and low prices over the past two years, these density maps are used to identify the relationships among the high and low gas price areas. Then, we use a data mining method called Apriori algorithm to address various questions such as whether the high priced stations tend to be spatially associated with low priced ones, and how such a pattern, if any, varies through time. We also explore the scale of these patterns using different spatial scales (e.g., size of the neighborhood) and temporal scales. Our preliminary results show that the co-location between the high and low price stations is a significant pattern at the hourly scale. Further analysis will be carried out to understand co-location patterns in different spatial scales and temporal scales.