I plan on using Processing, a java based programming language, to create a real time, visual representation of localized trending topics and tweet density within the city of Vancouver. This visualization is intended to be used as an interactive web app.
One feature of Twitter is that users are able to post tweets that contain a geotag. Geotagging associates the tweet with a longitude and latitude, but more recently has been updated to convert the location value into place names.
By plotting the geotagged tweets onto a map of Vancouver, I will cluster them by longitude and latitude into square kilometre wide blocks. The number of tweets per square block will be visualized through a 3D mesh overlay that will float above a map of the city. The height of the mesh overlay associated with a set area will be directly related to the number of tweets from that area within a set time frame.
For each area, the tweets will be parsed for common words or phrases which will be displayed either when the user hovers the mouse over a particular area, or in a fixed position over the mesh overlay. Ideally, these common words will give a brief overview of what’s happening in that area at that time, serving as localized trending topics.
For example, by examining the top tweeted words for a particular area, the user may be able to tell who’s performing at Rogers Arena that night, where there’s a traffic jam or what festivals are happening around the city.