Mining visitor interactions

This project leveraged Lag Sequential Analysis (LSA), a data mining technique on data representing the various events (micro and macro events as detected as change in system states recorded in the visitor interaction log files) that occur as visitor groups collectively interact with Connected Worlds museum exhibit. The lag sequential analysis works by distilling sequences by varying the temporal lags between consequent events. The method distilled thousands of sequences that were then labelled manually into three categories: (1) associations between human actions, (2) associations between the systemic events and (3) suggestive associations between human actions and systemic events. These included sequences with certain events preceding and succeeding critical states of the system which were difficult to identify during the visitor interaction.These also exposed a host of uninterpretable ``garbage'' associations (60% of the mined associations) that were the consequence of the way many unrelated events can co-occur within the system. This guided the innovation of the novel method to extract more meaningful and conceptual sequences.