Logging and Analyzing 360-Degree Video Users

User logging and analysis can be an important task in the domain of 360-degree video applications. This thesis studies what user logging in 360-degree video domain is, the reasons behind it, and how to do it efficiently.

The first contribution of the thesis is a definition of 360-degree video user logging, and a categorization for the different reasons for doing it. The study material used is scientific literature and company needs from five companies operating in the field. According to the thesis, the most popular reason for logging 360-degree video users is "viewport-adaptive streaming, transmission optimization, and saving bandwidth".

Despite the general benefits of log servers and databases, the related scientific implementations have not widely adopted them. Thus, the second contribution of the thesis is a reference architecture aimed at efficient distributed user logging and analysis. The reference architecture is used as a base for practical logging and analysis solutions. The solutions implemented are compared to those found in the related scientific literature. The comparison suggests that using publish-subscribe protocol in the logging of 360-degree videos users is a relatively novel approach, and the benefits of databases are also highlighted.

Different methods for user log analysis are studied with the emphasis on realtime visualizations. The third contribution of the thesis is real-time visualizations implemented using two approaches and the comparison of these to implementations made by other authors. The first approach is by drawing visualizations over 360- degree video during video playback and the second approach is by making traditional chart visualizations. The analysis needs of the five companies operating in the field are reflected against the visualizations implemented. A comparison to the state-of-theart related research reveals that real-time 360-degree video user analysis visualizations are not common.