Data Science is the current gold rush. While many industries have benefitted from applications of data science, including machine learning and Artificial Intelligence (AI), the applications in upstream oil and gas are still somewhat limited. Some examples of applications of AI include seismic interpretations, facility optimization, and data driven modeling – forecasting. While still naïve, we will explore cases where data science can be used in the day to day field optimization and development. The Midway Sunset (MWSS) field in San Joaquin Valley, California has over 100 years of history. The field was discovered in 19011 and had limited development through the 1960s. Since the start of thermal stimulation in 1964, the field has seen phased thermal flooding and cyclic stimulation. Recently there has been an increase in heat mining vertical and horizontal wells to tap the remaining hot oil. As with any brownfield, the sweet spots are long gone. Effort is now to optimize the field development and tap by-passed oil, thereby increasing recovery. The current operational focus includes field wide holistic review of remaining resource potential. Resources in the MWSS reservoirs are produced by cyclic steam method. Cyclic thermal stimulation has been effective as an overall depletion process and for stimulating the near wellbore region to increase production. It is imperative to properly identify target wells and sands for cyclic stimulation. Cyclic steaming in depleted zones or cold reservoirs is often uneconomical. The benefit comes when we can identify and stimulate only the warm oil. Identification of warm oil and short listing the wells for cyclic stimulation is a labor-intensive process. The volume of data can get so large that it may not be feasible for a professional to effectively do the analysis. In this paper, we present a case study of data analytics for high grading wells for cyclic stimulation. This method utilizes the machine power to integrate reservoir, and production data to identify and rank wells for cyclic stimulation and potentially increase success rate by minimizing suboptimal cyclic candidates.