Recently I have been rethinking the approach for my thesis because of the concerns I expressed before. To sum up, my previous focus was on the user’s interaction and on the communication process. To offer support in this context I was planning to use a device able to track constantly the user, with a certain degree of resolution.
As it turned out in the implementation we are working on, this was not possible due to the technological constraints of the positioning technique we choose and the integration in the mobile platform.
Therefore, here we are, with a mobile system that allows the user to attach a text message to a shared map. Nothing more. What we miss:
– the background of the user;
– the continuous tracking of his/her movements in the space (we have in fact a discontinuous information);
– a precise task on which to orientate the support system.
For these reasons I proposed to refocus the thesis on the clustering of the messages in space: namely on the content rather than on the user. This is based on a simple idea: usually information visualisation research proceed from the clustering of the words composing a certain amount of text, to the visualization of this clusters. Here I want to re-orientate the approach the other way around: we have a certain graphical distribution of messages, which is subsequent to a certain context pertinence and we want to reconstruct the topic descriptor of that context that generated the aggregation of messages.
The research question can be reformulated as:
Q_ How can we define the context of a place analysing the shared annotations that people do on the representation of that place, in such a way that this context is maximally useful for the information retrieval and the fruition of the information by the same users?