Wikipedia talk (ja)
This is the communication network of the Japanese Wikipedia. Nodes represent
users, and an edge from user A to user B denotes that user A wrote a message on
the talk page of user B at a certain timestamp.
Metadata
Statistics
Size  n =  397,635

Volume  m =  1,031,378

Unique edge count  m̿ =  656,330

Wedge count  s =  18,963,253,905

Triangle count  t =  251,868

Square count  q =  1,272,886,511

4Tour count  T_{4} =  86,037,333,232

Maximum degree  d_{max} =  170,852

Maximum outdegree  d^{+}_{max} =  170,851

Maximum indegree  d^{−}_{max} =  3,633

Average degree  d =  5.187 56

Average edge multiplicity  m̃ =  1.571 43

Size of LCC  N =  394,528

Diameter  δ =  10

50Percentile effective diameter  δ_{0.5} =  3.188 88

90Percentile effective diameter  δ_{0.9} =  3.862 74

Median distance  δ_{M} =  4

Mean distance  δ_{m} =  3.380 54

Tail power law exponent  γ_{t} =  1.921 00

Degree assortativity  ρ =  −0.281 185

Degree assortativity pvalue  p_{ρ} =  0.000 00

Clustering coefficient  c =  3.984 57 × 10^{−5}

Spectral norm  α =  2,914.80

Operator 2norm  ν =  1,464.24

Algebraic connectivity  a =  0.046 582 3

Nonbipartivity  b_{A} =  0.827 669

Normalized nonbipartivity  b_{N} =  0.027 208 6

Algebraic nonbipartivity  χ =  0.053 922 5

Spectral bipartite frustration  b_{K} =  0.004 189 78

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]

Jun Sun, Jérôme Kunegis, and Steffen Staab.
Predicting user roles in social networks using transfer learning with
feature transformation.
In Proc. ICDM Workshop on Data Min. in Netw., 2016.
