Twitter Orographies is an experiment, which tries to visualize the discussions or conversations that are based around a certain keyword. Picking up on conversations that take place on Twitter, it shows the different keywords that pop up, the importance of each of these keywords and how exactly they are related to each other in the conversations that ensue.
How this experiment works is really simple. Twitter Orographies keeps track of all streams of tweets, in real time, that contains the particular keywords that have been selected for visualization in the experiment. Even hashtags may be used in place of or as a substitute for keywords in order to filter out streams of tweets for the experiment. These tweets are then analyzed one by one so that they are more in a format that is expected of for the research. This is done by eliminating the common words whose semantic content are weak, out of the conversation – such as articles or prepositions, and adding the remaining words to a weighted network where nodes symbolize the topics of discussion.
The edges’ weights denote up to what level the topics are related to each other. These topics are then filtered depending on various conditions such as how often they are repeated and if they tend to pop up with the new topics. Old topics that do not reappear are removed. The final representation displays the conversation in form of a landscape, with the relevant topics forming the higher hills and new topics, the lower hills. Relatedness is depicted by distance and demonstrated by visual links.