The Semantic Thermometer is a playful, visual way to display current local temperature data. The program takes a location and accesses up-to-the-minute weather information for that location, using Yahoo’s Weather API and a Where-On-Earth Id. The current temperature reading for that location is translated from a number to a corresponding descriptive word (for example 77 degrees Fahrenheit would be translated as “warm” and 24 degrees Fahrenheit would be “frigid”) from a set of words I defined. The program sends the word to the Flickr API, where it searches for and retrieves the latest user-generated photos tagged with that word. These photos are displayed in a grid on the screen, with a new photo added every few seconds to show the current local temperature in images. When the user rolls over an image, the descriptive word appears. The Semantic Thermometer invents surprising and baffling new metaphors, as it works in the gap between a specific physical experience and the imprecise, complex language we use to describe it.