A Google+ (now defunct) dataset of 102,100 nodes and 13,548,655 directed edges across 132 ego networks with ground-truth circles, collected by McAuley and Leskovec (Stanford, 2012) from public Google+ profiles. Each record contains directed follow edges, ground-truth circle memberships (Google+ circles shared publicly), anonymized node feature vectors representing profile attributes. 133 Google+ ego networks with publicly shared ground-truth circles from users who voluntarily released their circle data, providing a uniquely clean community ground-truth from a directed social platform. Suitable for overlapping community detection, ego network analysis, directed social circle inference, link prediction.
A Twitter (ego-networks) dataset of 81,306 nodes and 1,768,149 edges, collected by McAuley & Leskovec via Twitter public crawl (2012). Each record contains directed user-user follow edges and ego-network circles. The dataset is notable for its directed ego-network dataset with ground-truth social circles. Intended for research into social network analysis, influence propagation, and community detection.
A Facebook (ego-networks) dataset of 4,039 nodes and 88,234 edges, collected by McAuley & Leskovec via Facebook app survey (2012). Each record contains user-user friendship edges and ego-network circles. classic ego-network dataset with ground-truth social circles. Suitable for social network analysis, community detection, and circle prediction.
Google
Twitter (X)
Facebook