This is the second part of an old project that I recently replicated in Python, it was presented as my Conclusion Work of my Post-Graduation at FIA School. This study is focused on my Facebook friends data and their friendship connections, collected through Facebook Netvizz App. This app treats json returns from Facebook Graph API, and formats in two files of nodes (friends) and edges (friendship relations).

I used the Python package Networkx to detect communities based on nodes relationships and to compute centrality measures, in order to understand patterns on detected communities and identify aspects of leadership and popularity in my personal friends. The graph of my social network and detected communities is illustrated in following.

In the following figure, we can observe leadership aspects on my social network.

Finally, we used these centrality measures and communities, as input features in a classification problem: predict who would be invited to my Wedding which append years later. The model reached an AUC of 0.86.

The Colab Notebook of this study is available here.

The data sets are available here.

• Social Network Analysis
• Community Detection
• Centrality Measures
• Facebook
• Social Network