DISCLAIMER: This repository is under development, with only some parts of it fully working.
Suite of Python and MATLAB files for analysis and forecasting of catalan and spanish elections. Automatically retrieves election results and polls from the internet (with e.g. Pandas), combines them and performs predictions on the expected evolution and number of sits each party will get.
Different techniques are or will be implemented:
- Correcting for bias and scaling.
- Hidden Markov model (polls result/actual voting)
- Kalman filter (assuming a sort of Wiener process)
- Vector ARMA models.
Includes number of sits/province and d'Hondt's law, which is applied in Spain and Catalonia.