The advent, in recent years, of less invasive and cheaper technologies for the detection of actions performed by a human in various fields, has led to a huge growth of interest in the analysis of human behavior. In particular, through the large amount of data obtained by means of these technologies and thanks to the increasing computational power available, it has been possible to analyze more and more in detail both indoor and outdoor human behavior. This thesis work is placed right in this context. Through the use of different types of sensors for data collection, we focused on the analysis of pedestrian behavior both outdoors (particularly in the context of a natural park) and in-doors. With regard to the latter analysis, the retail environment was taken into consideration; therefore, a series of strategies were developed for data collection in this context and for the analysis of consumer behavior with different levels of detail. Always starting from the trajectory data, the shopper’s habits have been analyzed at store, shelf and finally person level through the sentiment analysis. The entire analysis was conducted taking into consideration real case studies and consequently real data.
L'avvento, negli ultimi anni, di tecnologie meno invasive e sempre più economiche per la rilevazione delle azioni compiute da un essere umano in diversi campi, ha portato ad una enorme crescita di interesse relativamente al tema dell'analisi del comportamento umano. In particolare, attraverso la grande mole di dati ottenuti per mezzo di queste tecnologie e grazie alla maggiore potenza di calcolo disponibile, è stato possibile analizzare sempre più in dettaglio il comportamento umano sia in ambito indoor che outdoor. Il presente lavoro di tesi si colloca proprio in questo contesto. Attraverso l'utilizzo di diversi tipi di sensori per la raccolta dati, ci si è concentrati sull'analisi del comportamento dei pedoni sia all'aperto (in particolare nel contesto di un parco naturale) che al chiuso. Per quanto riguarda quest'ultima analisi, è stato preso in considerazione il contesto retail; sono state quindi sviluppate una serie di strategie per la raccolta dati in questo contesto e per l'analisi del comportamento dei consumatori su diversi livelli di dettaglio. Sempre partendo dal dato traiettoria, le abitudini dello shopper sono state analizzate a livello di negozio, di scaffale e infine di persona attraverso la sentiment analysis. L'intera analisi è stata condotta prendendo in considerazione casi di studio reali e di conseguenza dati reali.
HUMAN BEHAVIOUR ANALYSIS IN INDOOR AND OUTDOOR ENVIRONMENTS AND CLUSTERING OF TRAVELLED TRAJECTORIES / Vaira, Raffaele. - (2021 Oct 31).
HUMAN BEHAVIOUR ANALYSIS IN INDOOR AND OUTDOOR ENVIRONMENTS AND CLUSTERING OF TRAVELLED TRAJECTORIES
VAIRA, RAFFAELE
2021-10-31
Abstract
The advent, in recent years, of less invasive and cheaper technologies for the detection of actions performed by a human in various fields, has led to a huge growth of interest in the analysis of human behavior. In particular, through the large amount of data obtained by means of these technologies and thanks to the increasing computational power available, it has been possible to analyze more and more in detail both indoor and outdoor human behavior. This thesis work is placed right in this context. Through the use of different types of sensors for data collection, we focused on the analysis of pedestrian behavior both outdoors (particularly in the context of a natural park) and in-doors. With regard to the latter analysis, the retail environment was taken into consideration; therefore, a series of strategies were developed for data collection in this context and for the analysis of consumer behavior with different levels of detail. Always starting from the trajectory data, the shopper’s habits have been analyzed at store, shelf and finally person level through the sentiment analysis. The entire analysis was conducted taking into consideration real case studies and consequently real data.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.