Purpose-Nowadays, social networks, online media sharing and e-commerce platforms generate a vast amount of data, which, among other information, capture the interactions among the users. Storing, analyzing and exploiting the aforementioned information allow the exploration of hidden and unstructured patterns.
Design/methodology/approach-The associations among the users during their visit in a platform construct a graph network which capture all the generated data. Graph Neural Networks are applied in these data models, to make suggestions based on their topology. In the presented research, Graph Databases and Graph Neural Networks are utilized for data exploration and analysis in graph databases networks.
Findings-In this study, we compare the use of graph databases with relational databases for large-scale databases and we present that the use of graph neural networks over graph databases can be used efficiently to apply machine learning tasks for those datasets.
Originality/value-Thus, in this paper, we present the benefits of applying graph neural networks and graph databases for data analysis in large-scale data from social networks. Also, we examine to the efficiency of using graph databases over relational databases for analyzing those networks.