MathOverflow
These are interactions from the StackExchange site "MathOverflow". The network
is between users, and directed edges represent three types of interactions:
answering a question of another user, commenting on another user's question,
and commenting on another user's answer. The network is temporal. This
network is part of a series of network from multiple StackExchange sites.
Metadata
Statistics
Size  n =  24,818

Volume  m =  506,550

Unique edge count  m̿ =  239,978

Loop count  l =  116,109

Wedge count  s =  54,999,653

Claw count  z =  31,659,979,718

Cross count  x =  12,853,869,969,792

Triangle count  t =  1,403,919

Square count  q =  185,375,938

4Tour count  T_{4} =  1,703,382,088

Maximum degree  d_{max} =  11,309

Maximum outdegree  d^{+}_{max} =  5,931

Maximum indegree  d^{−}_{max} =  5,378

Average degree  d =  40.821 2

Fill  p =  0.000 389 617

Average edge multiplicity  m̃ =  2.110 82

Size of LCC  N =  24,668

Size of LSCC  N_{s} =  13,095

Relative size of LSCC  N^{r}_{s} =  0.527 641

Diameter  δ =  9

50Percentile effective diameter  δ_{0.5} =  2.652 69

90Percentile effective diameter  δ_{0.9} =  3.668 56

Mean distance  δ_{m} =  3.208 63

Gini coefficient  G =  0.854 886

Balanced inequality ratio  P =  0.143 503

Outdegree balanced inequality ratio  P_{+} =  0.120 448

Indegree balanced inequality ratio  P_{−} =  0.177 469

Relative edge distribution entropy  H_{er} =  0.832 983

Power law exponent  γ =  1.682 16

Tail power law exponent  γ_{t} =  1.871 00

Degree assortativity  ρ =  −0.215 181

Degree assortativity pvalue  p_{ρ} =  0.000 00

In/outdegree correlation  ρ^{±} =  +0.796 549

Clustering coefficient  c =  0.076 577 9

Spectral norm  α =  4,010.92

Operator 2norm  ν =  2,011.89

Cyclic eigenvalue  π =  1,997.09

Algebraic connectivity  a =  0.207 313

Reciprocity  y =  0.383 356

Nonbipartivity  b_{A} =  0.994 978

Normalized nonbipartivity  b_{N} =  0.132 313

Spectral bipartite frustration  b_{K} =  0.003 028 19

Plots
Matrix decompositions plots
Downloads
References
[1]

Jérôme Kunegis.
KONECT – The Koblenz Network Collection.
In Proc. Int. Conf. on World Wide Web Companion, pages
1343–1350, 2013.
[ http ]

[2]

Jure Leskovec.
Stanford Network Analysis Project.
http://snap.stanford.edu/, September 2014.
