Dataset

Directors

Firm-director matched records with synthetic regressors and outcome.

Bipartite incidence pattern for Directors

Graph Summary

A difficult real graph with slow convergence in the old benchmarks.

The primary benchmark graph is built from unique (id1, id2) pairs. The figure shows a binned sparsity pattern of the bipartite incidence block, with both partitions relabeled to contiguous integer identifiers. This block is the off-diagonal part of the corresponding graph Laplacian.

525,012rows
524,819unique edges
395,925id1 levels
400,850id2 levels
281,627components
796,775nodes

Variables

Columns in the clean CSV:

id1 id2 x1 x2 y

The v1 graph uses id1 and id2. No additional identifier-like columns are included in this clean v1 CSV.

Source Notes

Firm-director matched records based on SUNAT data collected by the author.

Identifiers are anonymized; regressors and outcome are synthetic benchmark variables.

Historical Benchmark

2017 SEC benchmark timings for this dataset, in seconds:

MethodCitationSeconds
MAP-Aitken (Guimaraes 2012) 688.8
MAP-SD (Gaure 2013) 215.6
MAP-CG-Sym (Correia 2016) 258.6
MAP+Prune (Correia 2016) 240.7
LSMR (Gomez 2016) 894.1