This paper proposes the use of spatial analysis of stimulation treatments and historical oil production to improve the selection of future candidate wells for workover treatments. Oil production data from four mature leases of the San Andres carbonate reservoir in the West Texas Permian Basin are analyzed to identify a potential correlation between historical oil production and the oil production improvement from well stimulation. Spatial grouping of producers is used to eliminate the bias introduced into the data because of workover failure, errors in production data, ongoing water and CO2 floods and other factors affecting the two correlated parameters. The spatial grouping approach is further extended to a semi-automated method for large datasets using the meanshift algorithm. To increase the robustness of the results, the analysis is repeated with ordinary kriging as an independent method. The results obtained from both methods confirm the presence of a positive correlation between the historical oil production and the well response after the stimulation. The results are also compared with porosity-height maps, indicating that wells located in zones with better porosity-height have better stimulation response. This has important implications in identifying wells with greater potential for oil production improvement from well stimulation. While the success of the treatment strongly depends on the selection, design, and implementation of the appropriate stimulation method, priority should be given to the wells that exhibit high historical oil production, or the wells located in highly productive zones. This practice will statistically increase the likelihood of maximized workover benefits. The presented analysis used the oil production data from the San Andres carbonate reservoir; however, the proposed method is applicable to gas reservoirs and unconventional reservoirs.