In this dissertation, I present a series of visual analytics tools designed to provide insights into tennis matches for the purpose of player improvement. These include analytic tools that can be divided into two basic categories: non-spatial-based, and spatial-based. The non-spatial-based tools acknowledge the difficulty in obtaining ball and player location tracking data and instead focus on tennis semantic data that is easier to collect, including scoring information, who is serving, and point outcomes (i.e., ace, double-fault, unforced error, forced error, and winner). Data collection techniques easily implementable by non-professional tennis players are outlined. This data is then combined using a set of novel, interactive visualizations whose utility is vetted through a small user study. Spatial-based tools are then proposed that bring to light the potential insights available when we can get reliable player and ball tracking data integrated with domain-specific semantic data. Although spatial data is still typically hard to come by outside of all but the major professional tennis tournaments, the case will be made that spatial-based tracking systems are beginning to find their way into college tennis programs and even into private tennis clubs. Research tools and efforts are described that show how we can use visual analytics techniques to provide players and coaches with meaningful, strategic insights into players' strengths and weaknesses.