The CDC diagnoses millions of cases of infectious diseases annually with observed disease curves peaking around mid-December and lulling in August and September. While this provides an accurate depiction of disease spread, its compilation takes too long for up-to-date monitoring. The ability to generate real-time disease distributions is important in identifying outbreaks and facilitating instant communication between authorities and health-care providers. We have attempted to characterize disease propagation using Twitter, expanding upon Google’s 2008 Flu Trends project. Our novel contribution is the development of a pipeline based model incorporating natural language processing and machine learning. The correlation coefficient between the Twitter disease distribution obtained via our approach and CDC data was 0.98. Our model further identified disease outbreaks that were not prevalent in the CDC distribution such as the parotitis outbreak in late 2014 that large hospital samples failed to identify. We additionally develop a differential equation based disease simulation (known as SEIR) in order to further validate our Twitter disease distribution model. Our model has the potential to greatly assist in the creation of an early-warning infection system by identifying disease outbreaks in real-time using the ever-growing social media sphere, representing a unique and powerful benefit to society.