如果NetworkX不好用，pip uninstall 然后install
1) 看每个node的social情况
In [1]: import networkx as nx In [2]: nx.read<点击TAB看选项> 比如 nx.read_adjlist nx.read_dot nx.read_edgelist 这个时候应该在有一系列edges和dot的文件夹 fbdata里 In [3]: g = nx.read_edgelist(‘0.edges‘) 此时g继承了read的各种method，查看g的情况 In [5]: len(g.nodes()), len(g.edges()) Out[5]: (333, 2519)
The 0 ego user then appears to have 333 friends（nodes）, and there are 2519 connections(edges) between these friends.
2） 看social network的structure
sg = nx.connected_component_subgraphs(g) [len(s) for s in sg] 这里因为sg是个generator object 书上的用法不对，但是我不知道怎么改 s1 =list(sg) s1 = s1[0] nx.radius(sg), nx.diameter(sg) 可能是版本问题，
ipython 学习笔记 2 network graph--NetworkX
原文：http://www.cnblogs.com/jiajiaxingxing/p/4548813.html