The concept of percent is ubiquitous in modern society. Percent is a common component across all forms of communication and occurs along a continuum from small scale (e.g. the tip after a meal) to large scale (e.g. deciding to evacuate from a natural disaster) decisions. Yet, the concept of percent is not simple. Percent is a functional number and therefore takes its conceptual meaning from its functional context. This dissertation analyses the concept of percent as a complex mathematical tool using interdisciplinary approaches focused on three key aspects: construction, context, and comparison. For construction, this work focuses on the inability of the concept of percent to coalesce as an integration of a network of ideas created from historical commercial practices and abstract numerical structures over the ages. For context, natural language processing applications and machine learning approaches are fine-tuned on a corpus of natural language and show the role of language formality as it pertains to percent as a functional number in comparative situations of many types. For comparison, the concept of percent is shown to be most similar to decimals, over integers and fractions, via a continuous-measures behavioral magnitude comparison study. These three interdisciplinary approaches present qualitative and quantitative insights relevant to addressing the known challenges of the concept of percent.