       Yaser Ghanam and Sheelagh Carpendale Department of Computer Science University of Calgary, Canada T2N 1N4 {yghanam, sheelagh}@ucalgary.ca  This report presents ChatVis as a 2,800 LOC tool for visualizing activities in instant messaging conversations. The tool visualizes at four different levels, starting from the conversation level down to the message level, the word level and finally the character level. The tool provides means to visualize the typing speed and rhythms of an arbitrary number of conversers. Additionally, it attempts to visualize intangible characteristic of a typed message such as hesitance (uncertainty) and emphasis. ChatVis allows for mouse as well as keyboard interactions to give the user a variety of filtering and zooming options. The system state is always saved allowing for animation through temporal aspects of the conversation. Moreover, the conversation can be persisted to XML and loaded back on demand for comparison purposes amongst different people.   The main purpose of this visualization, hereafter called ChatVis, is to augment activities in instant messaging conversations. This augmentation is attempted by including evidences of the theme, the tone and the length of the conversation as well as some manifestations of the personalities of conversers. The goal behind ChatVis is to make it possible for an observer to have a quick insight of what the conversation was all about, how long it lasted and how each of the conversers behaved in the conversation in terms of emphasis, hesitance, patience as well as typing speed and rhythm. Figure 1 shows the output generated by ChatVis to visualize a sample conversation between User1 and User2. Later in this paper, every component of this visualization will be explained in detail. This visualization may look like CrystalChat [1], but the purpose of ChatVis, data mapping and visualization variables are completely different. While CrystalChat focuses on the personal chat history, ChatVis is interested in the visualizing instant messaging at the conversation level.