Anonymised tourist-by-site rating matrix (17 x 80) used in the bipartite motif audit

Anonymised tourist-by-site rating matrix (17 tourists x 80 sightseeing sites, ratings 0 to 5) underlying the tourism bipartite network first reported in Shao, Ieiri and Hishiyama (2021), "Discovering Multiple Clusters of Sightseeing Spots to Improve Tourist Satisfaction Using Network Motifs", IEICE Transactions on Information and Systems E104.D(10):1640-1650, doi:10.1587/transinf.2020EDP7258. Tourist identifiers are replaced by T1 to T17 and site names by codes; no personal data are included. Placing an edge wherever a rating is greater than zero gives the 17 x 80 bipartite network with 637 edges analysed in the author's re-analysis note "Auditing bipartite motif interpretations: a worked example with conservation checks and open-path decomposition" (submitted to PeerJ Computer Science, 2026), which this record accompanies.

Files: tourism_matrix.csv (the matrix; first column tourist id, header row site codes), README.md (provenance, anonymisation, network construction, citation), LICENSE (CC BY 4.0), and tourism_deposit_v1.zip bundling the three. The matrix is byte-identical to the file read by tourism_overlap_null.py in the code deposit under concept DOI 10.5281/zenodo.21122972 (version 5: 10.5281/zenodo.22287908); the degree-preserving null tests use attrimotif 1.1.1 (10.5281/zenodo.22259239).

Version 1.1 (2026-09-04). README corrected: the matrix was collected in a Kyoto walk-rally experiment with 17 tourists reported in Ieiri, Nakajima and Hishiyama (2018), IEICE Transactions on Information and Systems (Japanese edition) J101-D(9):1325-1333 (in Japanese), and reused in Shao, Ieiri and Hishiyama (2021); evaluations were collected for visited spots, so a value of 0 means the spot was not visited. The depositing author confirmed with the authors of the data-collection study that the consent obtained from participants in that study covers public re-use of this de-identified matrix, and the co-authors agreed to the deposit; the site-code key is held by the authors of the data-collection study and is not released. Data values are unchanged from version 1.0.

Version 1.2 (2026-09-04). README wording corrected: the matrix is described as the note's only fully reconstructable dataset (not a second dataset) and a stale placeholder about a DOI assigned at publication was removed. Data values unchanged.

Ethics note (2026-09-04). Waseda University confirmed, at the depositing author's enquiry, that the 2018 data collection (Ieiri, Nakajima and Hishiyama 2018) did not require ethics review. The re-analysis note collected no new human-subject data.