briatte

Using open legislative data to map bill co-sponsorship networks in 15 countries

[Published at Opening Parliament.org](published at http://blog.openingparliament.org/post/113803051548/using-open-legislative-data-to-map-bill).

A few years back, Kamil Gregor published a post on the OpeningParliament.org blog under the title “Visualizing politics: Network analysis of bill sponsors”.

His post, which focused on the lower chamber of the Czech Parliament, showed how basic social network analysis can support the exploration of parliamentary work, by revealing the ties that Members of Parliament create between each other through the cosponsorship of private bills.

As Kamil observed, this kind of research is heavily dependent on open legislative data, which do not exist for many parliaments. There has been, however, some amount of progress in that area -- and even when parliaments do not develop open data _portals

briatte

library(ROAuth)

requestURL <- "https://api.twitter.com/oauth/request_token"
accessURL <- "https://api.twitter.com/oauth/access_token"
authURL <- "https://api.twitter.com/oauth/authorize"

# replace XYZ with your Twitter API consumer key
consumerKey <- "XYZ"

# replace ... with your Twitter API consumer secret

briatte

# introduction vite fait au t-test (comparaison de moyennes)

# le jeu de données fictif ci-dessous est à remplacer par
# data = read.csv("/chemin/vers/votre/fichier.csv", stringsAsFactors = FALSE)

data = data.frame(site = letters[ 1:15 ],
                  categorie = sample(LETTERS[ 1:2 ], 15, replace = TRUE),
                  mesure = runif(15))

# voilà à quoi doivent ressembler les données

briatte

library(dplyr)
library(ggplot2)
library(readr)
library(survey)

ess5 = read_csv("ESS5_FR_SDDF.csv") %>%
  left_join(read_csv("ESS5FR.csv"), by = c("cntry", "idno")) %>%
  ess_prepare

ess5 = svydesign(

briatte

library(Matrix)

## building a 729670 x 1167 object (fast enough)

M = list()

pb = txtProgressBar(min = 1, max = length(followers_m))

for(i in 1:length(followers_m)) {

briatte

library(dplyr)
library(rvest)

r = "http://www.apn.dz"

n <- function(h, g) {
  
  h = html_nodes(h, ".listing-summary")
  data_frame(
    id = html_attr(h, "data-link-id"),

briatte

R code to build the Icelandic legal code by connecting law articles through their cross-references.

The raw data can be loaded in R as an object called net of class network by loading this shortlink:

source("https://goo.gl/q1JFih")

Edges are weighted by the number of (directed) mentions. Nodes are sized by their unweighted degree centrality and colored by their period of introduction into the code.

briatte

# issue when assigning a vector of edge values:

# x %e% "y" = 1:z             ## does not assign 1:z, creates duplicates
# set.edge.value(x, "y", 1:z) ## does not assign 1:z, creates duplicates
# set.edge.attribute(x, "y", 1:z)     ## assigns 1:z, as expected

# using version 1.12.0
library(network)

# minimal reprex

briatte

### delete tex files in slides folder

cd ~/Documents/Teaching/SRQM/Course

rm *.aux *.dvi *.toc *.log *.nav *.out *.snm *.synctex.gz *.tex.bak *.vrb *.run.xml *-blx.bib *.idx *.fdb_latexmk

### delete log files in replication folder

cd ~/Documents/Teaching/SRQM/Replication/

briatte

/*
	Styling for Markdown preview in TextMate, copty to:
	~/Library/Application Support/TextMate/Managed/Bundles/Themes.tmbundle/Support/web-themes/SRQM/style.css
*/

@import url(http://fonts.googleapis.com/css?family=Ubuntu:400,500,400italic);
@import url(http://fonts.googleapis.com/css?family=Ubuntu+Mono);

*, p {
	font-family: "Ubuntu", sans-serif;