Those who depend on public support -- whether they be individuals such as politicians and celebrities, or institutions such as corporations and non-profits -- use increasingly sophisticated tactics to manage communications with the outside world. Many employ individuals and/or firms to craft and control the messages they send. All give careful consideration to the selection and arrangement of the words in their messages, knowing that subtle variation in language can lead to significant differences in the impressions they make and, consequently, the opinions they help shape. In some cases, communicators test these arrangements of words for effectiveness before deploying them in broader use. And once a phrase has been constructed and vetted, the communicating parties strive for consistency and repetition.

These tactics lead to a distinct, and we believe telling, corpus. Our project attempts to illuminate patterns, and perhaps strategies, of language use in one such corpus: White House communications intended to influence public opinion. Our goal is to develop a tool that allows interested parties to find and compare what they believe to be intriguing phrases, see when those phrases were introduced, and analyze how their frequency has changed over time. Additionally, we want our users to be able to construct, share and discuss their interpretations of the data. Key questions that we imagine our users attempting to answer include: Why was a phrase introduced when it was? Why is the use of certain phrases increasing or decreasing in a particular manner at various points in time? And why are some phrases more popular than others throughout time as well as within distinct periods of time? Our target persona is a user that holds an interest in politics and/or communications, as well as a slightly above average understanding of quantitative data analysis.

While we purposely avoid making hypotheses about the popularity and behavior of specific phrases, we do expect patterns in the behavior of phrases generally. We believe that most phrases have a life-span, likely corresponding to their ability to affect opinion. We hypothesize that successful phrases accelerate in use as they catch on, but eventually flatten out as prolonged saturation diminishes their impact. Additionally, we believe phrase use will be highly correlated to specific events. We hope our visualization does more than just reveal the occurrence of events; rather, we hope it helps make public the language strategy of the White House leading up, or in response, to specific events.