Voters changes in Czech elections 2013 → 2017

Estimates of numbers of people changing parties between Czech elections 2013-2017 based on ecological inference model.

Individual parties:

Example (i: 1-8)

index.html?i=3

Overview

See overview chart

Analysis flow

1. download data for 2017 (to /json): downloader.py
2. extract data (list_2017.csv): scraper.py
3. join 2013 and 2017, select random polling station only, to be able to run R calculation (data_filtered.csv and data_filtered_random.csv): join_data.py
4. analysis - ecological inference (matn_filtered.csv, matp_filtered.csv): ei_psp.py calling ei.r
5. manually recalculate numbers to real total matn_filtered_adjusted.csv
6. manually reorder parties to have better matrix matn_filtered_adjusted_reordered.csv
7. manually add labels/party names matn_chart.csv
8. manually prepare lists of parties list_2013_filtered.csv and list_2017_filtered.csv
9. overview chart: bubbles_psp.html
10. calculate errors of model presnost.ods
11. individual party charts (with ?i=1 parameter, i: 1-8) index.html and index_cs.html
12. generate pictures (PNGs): generate.py

CC-BY Michal Škop