Location has an impact on business success and comprehensive location analysis is critical for location decision making. Based on agglomeration and competition effects, performance-related pattern analysis is helpful for location selection. However, few studies have investigated whether the performance-related patterns vary with different regions. In this paper, we explored the restaurants’ performance-related characteristics in different areas using Getis-Ord Gi* statistic. We found that there is a performance-pattern difference of hot/cold clusters between different regions. For the high and low revisit intention region, there are statistically significant clusters of restaurants with high revisit intention ("hot" clusters) in the northeast and west, as well as a cold cluster in the center. Nevertheless, there are no cold or hot clusters in the medium revisit intention region. The results of this study are helpful for deep understanding of restaurant performance in terms of different locations and have consulting value for restaurants on siting problem.