Can I work on a model that will score a body of text as being overtly political, similar to how sentiment analysis works? The goal would be to provide a tool for authors to generate text that performs equally across a diverse cross section of the population regardless of age, location, political affiliation, etc.