/parliament-viz.ch

Swiss parliament visualization project

Primary LanguageJupyter NotebookMIT LicenseMIT

Logo

parliament-viz.ch

Suggestion of a visualization interface for Swiss parliament deputies, following a complete Data Visualization pipeline, from scraping until visualization.

Final project of Applied Data Analysis course (ADA CS-401 EPFL 2016) taught by Michele Catasta.

Gael Lederrey, Jonas Racine and Joachim Muth

Find the final visualization at www.parliament-viz.ch

Similar Project

A project about predictions on the same data exists here.

Both teams collaborated on various aspects of the project, including the scraping and the understanding of the data.

Abstract

The Freedom of Information Act came into force on 1 July 2006. It is intended to promote transparency with regard to the purpose, organisation and activities of the federal administration, while guaranteeing access to official documents produced after 1 July 2006.

The motivation for this project comes from the gulf existing between the political community and citizens. However, it represents a large number of hardly understandable documents. We suggest a visualization of what could be a light and user-friendly public interface of Swiss parliament activities.

Data description

The whole data schema is accessible through metadata web page and can be explored using Pragmatica WOData visualization tool

We provide here a short summary (when the name is not self-explaining) of the main tables we consider in the project.

  • Transcript contains textual transcription of all speeches made during regular parlamentary session.

  • Business contains initiatives (motions, postulates, interpellation, ordinary questions, questions) made by deputies.

  • BusinessRole contains role of each deputy in initiatives (author, cosignee, ...)

  • BusinessStatus

  • BusinessType

  • MemberCoucil lists all parliament deputies

  • MemberParty

  • Party

  • Session contains the date of each parliamentary session

  • Tags classifies the initiatives into themes

Folder hierachy

  • graph: all computation of JSON for d3.js visualization
  • logo : a very mysterious folder
  • nlp: NLP tool used for previous speech theme extraction, not used in final delivery
  • poster: latex files for poster presentation
  • scraping: scrap and analyse of data
    • scraper.py: scraper class to retrieve database table
    • tables.py: class to reproduce tables used in visualization
  • viz: website files

Helps and others

Parliamentary Terms

Complicated and precise terminology, specific to the political area. Lexicon of Parliamentary Terms

Data-Driven Documents

The D3 javaScript library will allow us to design a clean and interactive visualization. D3 website

Licences

Open Data / Web services statement of Parliament website The provided information can be freely used, under reasonable conditions (no alterations, source indicated, date of download indicated).

Deliverable

The final product takes the form of an online website presenting an interactive visualization of Swiss Parliament deputies activity. This website will be on parliament-viz.ch.

This visualization on:

  • the d3.js library
  • jQuery
  • bootstrap

Visitors have a general overview of the parliament where they can fly over each point and see some information about the deputy. Additionnaly, the whole parliament recolor itself to show the partnership the deputy maintains with the others.

A second visualization can seen by clicking the "Clustering" button, where the visitor can group the deputies on some criterias, as well as color them.

The website contains more information on page About

Chronology

  • October 26th, 2016: Start of the project
  • November 6th, 2016: Project approval
  • November 9th, 2016: Developing Python scraper
  • November 17th, 2016: Developing Python NLP analyser
  • December 1st, 2016: First cluster visualization
  • December 6th, 2016: Phone appointment with Alain Rebetez
    • Discussion about the paliamentary system
    • Discussion about what kind of informations could interest a journalist
    • Reorientation of the project, more on visualization, less on interpretation
  • December 8th, 2016: Abandon of NLP tool for theme extraction, according with reorientation
  • January 3rd, 2016: First "parliament-overview" visualization
  • January 6th, 2017: Meeting with Philippe Nantermod
    • Discussion about the paliamentary system
    • Demonstration of the visualization and how intuitive it looks like for an external user
    • New ideas about how extract partnerships and interests of deputies
  • January 20th, 2017: Partnerships based on cosigned initiatives
  • January 23th, 2017: Merge of the two visualization: "parliament-overview" and "clusters"
  • January 24th, 2017: Interests based on theme of redacted initiatives
  • January 28th, 2017: Sub-clusters and majority by moving clusters into the visualization
  • January 29th, 2017: Website
  • January 30th, 2017: Poster
  • January 31th, 2017: Project presentation at Applied Machine Learning Days (EPFL)

Acknowledgements

A warm thank you to Alain Rebetez, political journalist, Philippe Nantermod, PLR deputy at National Council, and Mathias Reynard, PS deputy, for their expertise in Swiss parliamentary system, their advice and their external point of view about the project.