R/data_irps.R
irps_corruption.RdThis network reconstructs a Czech political corruption case, known as the Rath affair, from publicly available archival data. David Rath was a Social Democrat politician who served as Minister of Health and then as governor of the Central Bohemian region. The police arrested him on 14 May 2012 as he carried a wine box that held seven million Czech crowns in cash. He and his associates manipulated public tenders in the region, several of them funded by the European Union, and took bribes from the firms that won them. The courts convicted Rath and most of his co-defendants, and he began a seven year prison sentence in 2019.
The network contains 11 actors connected through three binary, undirected layers. The "collaboration" layer records communication and tasks that two actors carried out together. The "transfers" layer records bribes and other transfers of resources. The "preexisting" layer records the ties that predate the affair, such as kinship, friendship, and shared political or professional affiliations.
The network also includes two node attributes.
politician marks the actors who held political office,
and gender records whether an actor is "male" or "female".
Diviak and colleagues use these data to show that the network is small, sparse, and centralised on Rath and two other actors, and that the preexisting ties underpin the criminal ones. This is a trade-off that covert networks face: ties that are already there are safe to use, but they also concentrate the network on a few actors and make it easier to disrupt.
data(irps_corruption)#> -- # Czech Rath affair corruption network --------------------------------------
#> # A labelled, multiplex, undirected network of 11 actors and 24 collaboration
#> ties, 15 transfer ties, and 5 preexisting ties
#>
#> -- Nodes
#> # A tibble: 11 x 3
#> label politician gender
#> <chr> <lgl> <chr>
#> 1 David Rath TRUE male
#> 2 Petr Kott TRUE male
#> 3 Lucia Novanska FALSE female
#> 4 Katerina Pancova TRUE female
#> 5 Pavel Drazdansky FALSE male
#> 6 Martin Jires FALSE male
#> # i 5 more rows
#>
#> -- Ties
#> # A tibble: 44 x 3
#> from to layer
#> <int> <int> <chr>
#> 1 1 2 collaboration
#> 2 1 4 collaboration
#> 3 1 7 collaboration
#> 4 1 8 collaboration
#> 5 2 3 collaboration
#> 6 2 4 collaboration
#> # i 38 more rows
#>The layer names shorten those that the article uses, so that each is
a single word: "transfers" for resource transfer, and "preexisting"
for pre-existing ties.
The two node attributes are coded 0/1 in the source data.
Here politician is logical, and gender is "male" for 0 and
"female" for 1.
The actor names are those that the court and media records use. Diacritics are removed from them, so "Drazdansky" appears here for "Drážďanský".
Diviák, Tomáš, Jan Kornelis Dijkstra, and Tom A. B. Snijders. 2019. "Structure, multiplexity, and centrality in a corruption network: The Czech Rath affair". Trends in Organized Crime 22(3): 274-297. doi:10.1007/s12117-018-9334-y