Awesome

This repository contains code to build cosponsorship networks from bills passed in the lower and upper chambers of the Italian Parliament.

• interactive demo
• static plots
• more countries

HOWTO

Replicate by running make.r in R. The data.r script will run two different scripts to download the sponsors from both chambers, followed by the bills (disegni di legge) that they have cosponsored in each legislature. The data for the lower chamber are collected via the SPARQL endpoint of its open data portal.

The data-ca.r script has been set to collect bills for legislatures 9-17 of the lower chamber. The start parameter found at the top of the script can be decreased to any value down to 1 to collect additional data.

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.

DATA

Bills

• legislature -- legislature number
• ref -- a reference of the form "Atto [Chamber: Camera/Senato] n. [number]"
• date -- date of introduction
• title -- short description
• authors -- first authors
• cosponsors -- cosponsors
• n_a -- number of first authors (see note below)
• n_c -- number of cosponsors

Sponsors

• legislature -- legislature number
• url -- sponsor URL, as used in the bills data
• name -- sponsor name
• sex -- gender (F/M)
• born -- year of birth
• constituency -- constituency, stored as the string to its Wikipedia Italiano entry
• party -- political party or parliamentary coalition
• committee - committee memberships, semicolon-separated
• nyears -- time in office before start of legislature (in years)
• photo -- file path to sponsor photo

NOTE

Italy has a strict distinction between first author(s) (prima firmatori) and cosponsor(s) (cofirmatori), and there is, in a limited number of cases from the Senate, multiple first authors, with or without cosponsors. This means that the cosponsorship graphs might contain two kinds of directed ties:

• mutual ties between first authors, and
• ties from cosponsors to first authors.

In order to keep things comparable across countries, it is probably best to treat only the first "first author" as such, and then treat all other sponsors as cosponsors. Adding mutual ties between first authors only marginally increases the density of the graphs.

THANKS

Thanks to Jeroen Ooms, who helped with a previous version of the code by providing a trick to work with the misconfigured cache server of the Senato website.