TripAdvisor

This is the rating network from TripAdvisor, where users rate individual hotels. We filtered out ratings for which the username was "lass=" (26,590 instances) and "A TripAdvisor Member" (37,704 instances), as these are probably anonymous ratings, or errors in the dataset.

Metadata

CodeTA
Internal namewang-tripadvisor
NameTripAdvisor
Data sourcehttp://times.cs.uiuc.edu/~wang296/Data/
AvailabilityDataset is available for download
Consistency checkDataset passed all tests
Category
Rating network
Dataset timestamp 2001-03-27 ⋯ 2009-01-11
Node meaningUser, hotel
Edge meaningRating
Network formatBipartite, undirected
Edge typeRatings, multiple edges
Temporal data Edges are annotated with timestamps
Snapshot Is a snapshot and likely to not contain all data

Statistics

Size n =147,075
Left size n1 =145,316
Right size n2 =1,759
Volume m =175,765
Unique edge count m̿ =175,655
Wedge count s =23,704,235
Claw count z =5,893,488,792
Cross count x =2,022,222,569,935
Square count q =11,327
4-Tour count T4 =95,259,242
Maximum degree dmax =2,138
Maximum left degree d1max =22
Maximum right degree d2max =2,138
Average degree d =2.390 14
Average left degree d1 =1.209 54
Average right degree d2 =99.923 3
Fill p =0.000 687 197
Average edge multiplicity m̃ =1.000 63
Size of LCC N =146,988
Diameter δ =14
50-Percentile effective diameter δ0.5 =5.478 03
90-Percentile effective diameter δ0.9 =7.153 99
Median distance δM =6
Mean distance δm =5.949 55
Gini coefficient G =0.572 841
Balanced inequality ratio P =0.291 614
Left balanced inequality ratio P1 =0.450 867
Right balanced inequality ratio P2 =0.317 480
Relative edge distribution entropy Her =0.848 532
Power law exponent γ =6.925 22
Tail power law exponent γt =2.981 00
Tail power law exponent with p γ3 =2.981 00
p-value p =0.000 00
Left tail power law exponent with p γ3,1 =3.311 00
Left p-value p1 =0.000 00
Right tail power law exponent with p γ3,2 =2.231 00
Right p-value p2 =0.000 00
Degree assortativity ρ =−0.099 365 7
Degree assortativity p-value pρ =0.000 00
Spectral norm α =28.736 5
Algebraic connectivity a =0.000 126 145
Spectral separation 1[A] / λ2[A]| =1.020 81
Negativity ζ =0.409 877
Algebraic conflict ξ =0.004 594 92
Spectral signed frustration φ =0.000 480 856
Controllability C =143,557
Relative controllability Cr =0.976 080

Plots

Fruchterman–Reingold graph drawing

Degree distribution

Cumulative degree distribution

Lorenz curve

Spectral distribution of the adjacency matrix

Spectral distribution of the normalized adjacency matrix

Spectral distribution of the Laplacian

Spectral graph drawing based on the adjacency matrix

Spectral graph drawing based on the Laplacian

Spectral graph drawing based on the normalized adjacency matrix

Degree assortativity

Zipf plot

Hop distribution

Double Laplacian graph drawing

Delaunay graph drawing

Item rating evolution

Edge weight/multiplicity distribution

Temporal distribution

Temporal hop distribution

Diameter/density evolution

Signed temporal distribution

Rating class evolution

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] Hongning Wang, Yue Lu, and Chengxiang Zhai. Latent aspect rating analysis on review text data: A rating regression approach. In Proc. Int. Conf. on Knowl. Discov. and Data Min., pages 783–792, 2010.
[3] Hongning Wang, Yue Lu, and ChengXiang Zhai. Latent aspect rating analysis without aspect keyword supervision. In Proc. Int. Conf. on Knowl. Discov. and Data Min., pages 618–626, 2011.