Skip to content

Latest commit

 

History

History
33 lines (21 loc) · 1.8 KB

README.md

File metadata and controls

33 lines (21 loc) · 1.8 KB
Raw

Parlamentsspiegel Scraper

The Parlamentsspiegel collects the federal parliamentary documentation of Germany.

Unfortunately, the documents on Parlamentsspiegel lack consistent meta data. For example, they use weird abbreviations for the German federal states, like SACA for Sachsen-Anhalt. That's far away from common standards like NUTS or ISO 3166... The translation between Parlamentsspiegel's weird country codes and more widely used codes can be found in input/lookup_laender_ps.csv

Also, the HTML structure has no well defined css classes, which makes the parsing a bit annoying.

The crawler is written in Python, the parsing in R.

Setup

  • create a virtual environment: python3 -m venv env
  • activate it: source env/bin/activate
  • install requirements: pip3 install -r requirements.txt

Use it

Define your search interest

  1. Write your search words for the search input field in input/searchwords.csv
  2. and the official tags ("Schlagworte") you're interested in input/keywords.csv

Run scripts

  1. 01_get_overview.py: to fetch all overview tables, which will be stored as html files in input/html/overview/ and relevant links in input/data/links_beratungsstand.csv
  2. 02_get_detailpages.py: to get all the information for every single document, the resulting html files can be found in input/html/beratungsstand/
  3. 03_parsing.R: to free metadata and to get a csv file with parsed information, that can be found in input/data/df.csv

To Do

  • clear distiction between input and output
  • automate folder creation
  • improve metadata parsing