Techniques used in yoga existed long before science as we know it. However, with Yoga westernization, the proliferation of scientific studies addressing the subject is exponential. Even though the literature presents thousands of related articles, many studies are considered flimsy - the diversity of methods, areas of interest, and focus cause it to become a complex environment without general guidelines for the researcher. This status may represent one of the impediments to the field's evolution. This study aims to elaborate a global structure of the literature on Yoga to open the door for multidisciplinary collaboration and bring new light to some relevant aspects. Due to its complexity, we understand that techniques to model complex domains are necessary. Contemporary methods of Artificial Intelligence have promoted advances in Bioinformatics, including text mining to scientific literature. Through the vector representation of texts, we got semantic abilities and could organize data in a way that allowed us to acquire a global vision of this literature. Additionally, we made the platforms available to Yoga researchers to enable them to evaluate our findings and make their forays into the literature. Besides better understanding why we study Yoga, we discovered that the literature offers insights into the relationships in broader aspects.