Riassunto analitico
Nowadays we are witnessing a continuous presence of mobile devices such as smart phones, laptop and smart watches, which have become a natural part of our daily life. The use of these objects and their unobtrusive presence in society has stimulated researchers to search for innovative applications fields that can help people in all aspects of life: from the most psy- chological ones to social and working conditions. This study consists of finding confirmation of particular methods to help people gain more productivity in the fields of work, using precisely those technological tools, now within everyone’s reach. We investigated for a possible relationship between features coming from smart phone usage and physiological data and we tried to use these features to automatically infer a person engagement during an activity.To do so we run 2 data collection: a pilot to get an overview of the data set and to understand the feasibility of this work and a study to explore the data and reach the goals of this thesis. Hence, for the study, we collected, after data cleaning, 361 data points from 9 workers for 2 weeks in situ data collection campaign. Our results show a correlation between physiological features and phone usage for the engagement level perceived by a worker during a task, in particular seems that physiological features of momentary sympathetic activation (phasic EDA component) are the most characterizing. More over we found that non engaged workers can be identified with high reliability. Using a Naive Bayes, we achieved a recall of 87% which is 35 and 38 percentage points higher than Dummy and RF. Overall, our findings may inform the design of a system that allows workers to self monitor their engagement status and therefor can act upon the obtain feedback.
|