Wikipedia talk (sr)
This is the communication network of the Serbian 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 =  103,068

Volume  m =  312,837

Unique edge count  m̿ =  153,042

Loop count  l =  17,756

Wedge count  s =  1,421,288,750

Claw count  z =  19,063,671,862,197

Cross count  x =  218,240,304,747,369,440

Triangle count  t =  103,105

Square count  q =  85,487,620

4Tour count  T_{4} =  6,369,340,732

Maximum degree  d_{max} =  47,687

Maximum outdegree  d^{+}_{max} =  47,685

Maximum indegree  d^{−}_{max} =  9,057

Average degree  d =  6.070 50

Fill  p =  1.440 66 × 10^{−5}

Average edge multiplicity  m̃ =  2.044 13

Size of LCC  N =  102,816

Size of LSCC  N_{s} =  2,416

Relative size of LSCC  N^{r}_{s} =  0.023 440 8

Diameter  δ =  7

50Percentile effective diameter  δ_{0.5} =  2.843 54

90Percentile effective diameter  δ_{0.9} =  3.791 96

Median distance  δ_{M} =  3

Mean distance  δ_{m} =  3.219 08

Gini coefficient  G =  0.820 749

Balanced inequality ratio  P =  0.149 006

Outdegree balanced inequality ratio  P_{+} =  0.056 579 0

Indegree balanced inequality ratio  P_{−} =  0.246 863

Relative edge distribution entropy  H_{er} =  0.681 060

Power law exponent  γ =  6.070 31

Tail power law exponent  γ_{t} =  2.831 00

Degree assortativity  ρ =  −0.330 145

Degree assortativity pvalue  p_{ρ} =  0.000 00

In/outdegree correlation  ρ^{±} =  +0.683 829

Clustering coefficient  c =  0.000 217 630

Directed clustering coefficient  c^{±} =  0.038 407 8

Spectral norm  α =  5,147.51

Operator 2norm  ν =  2,609.00

Cyclic eigenvalue  π =  2,550.56

Algebraic connectivity  a =  0.043 711 9

Spectral separation  λ_{1}[A] / λ_{2}[A] =  1.589 35

Reciprocity  y =  0.121 365

Nonbipartivity  b_{A} =  0.529 573

Normalized nonbipartivity  b_{N} =  0.011 641 0

Algebraic nonbipartivity  χ =  0.024 720 5

Spectral bipartite frustration  b_{K} =  0.002 193 01

Controllability  C =  99,966

Relative controllability  C_{r} =  0.969 903

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.
