Awesome

This repository contains code to build cosponsorship networks from bills passed in the Irish Parliament.

• interactive demo
• static plots
• more countries

HOWTO

Replicate by running make.r in R.

The data.r script downloads information on bills and sponsors, including their photos (which should all download fine). The build.r script then assembles the edge lists and plots the networks, with the help of a few routines coded into functions.r. Adjust the plot, gexf and mode parameters to skip the plots or to change the node placement algorithm.

The data are extremely sparse -- given how the Irish legislature works, and given the small number of parliamentarians in both chambers, there are only a handful of cosponsorships per year. As far as legislative cosponsorship goes, the Irish case is best understood as a borderline case that gives the minimal dimensions of a cosponsorship network.

DATA

Bills

• chamber -- lower (da) or upper (se)
• legislature -- legislature id
• ref -- bill id
• origin -- private (PM) or governmental (GOV) bill
• year -- date (yyyy-mm-dd)
• name -- title
• url -- URL
• authors -- URL to sponsors list
• sponsors -- semicolon-separated integer ids of sponsors
• n_au -- total number of sponsors

Sponsors

The saved version of the sponsors.csv dataset contains only url (profile URL, shortened to its numeric id), legislature and chamber (same as for bills). All other variables are collected on the way to plotting the networks:

• name -- name (duplicates solved by numbering them)
• born -- year of birth (int)
• photo -- photo URL, as a filename
• party -- party affiliation, abbreviated
• constituency -- constituency, stored as the string to its Wikipedia English entry
• sex -- gender (F/M), sometimes imputed from first and family names
• nyears -- seniority, in intervals of 5 years (computed on the fly)

Note -- chamber chairs (Cathaoirleach and Ceann Comhairle) are coded as "parties", even though they are not. Chairs do not show up in the networks.