An analysis of the impact that towny candidates had on the 2019 european elections in Italy, and some other fun stuff!
Do filler candidates influence the election on a national level?
2019 European election in italy (only local, no estero)
- a candidate belongs to a party
- circoscrizione > regione > provincia > comune
- all candidates belong to at least one circoscrizione
- a national candidate is a candidate which gets all the votes that expressed no preference in a certain circoscrizione
- are filler candidate (riempi-lista) is a candidate who is used by the party just to increase the number of preferences to the lista. ( QUANTITATIVE DEF:
- def 1, for a party: given the total number of preferences for a party, if a candidate has a low proportion of preferences party-wise q% (E.g. q=0.5), but on the total of preferences for the same party in a comune is > p% (E.g. p=50)then it is a filler candidate for a party.
- def 2, in general: a candidate which gets q% (E.g. q=70) of their preferences from comuni in which the same candidate gets p% (E.g. p=70) of the votes, is a filler candidate this def 2 is flawed )
[x] get the dataset in csv [x] organize dataset for python parsing [x] examine the dataset and clean it (17/08)
[X] how the preferences are spread out in a cumune by population [X] how the preferences are spread out in a cumune by population [] group by circoscrizione and highlight party GIOVANNI(21/08)
[] set percentages for filler candidates classification ALESSANDRO CON DEF1(21/08) [] consider big, small city effects [] Does their birthplace have any correlation with origin of votes? LORENZO(21/08)
[] scope of the party [] scope of the elections
[] where do succesful candidates get their votes [] Does their birthplace have any correlation with origin of votes? [] how candidates get elected in their party in relation with the votes they got (where they voted directly or where they pushed up by their party?) [x] how are votes distributed in a party? LORENZO [] how actually elected canddates distribute per party LORENZO (21/08 PRIORITA)
[] how representative are the elected candidates [] how polarized were the elections
-
majority of people did not express preferenze
-
for visualization I random sampled from the population weighting on the population of the settlement, correlation was kept at the same values roughly + some variation
-
there seems to be no correlation between area/population/density and homogeneousness of votes perhaps there is too much noise, clustering by parties might make this better
-
how much compromise must parties have to reach a high consensus? depending on the circoscrizione this number changes.
-
checking how polarized elections are in different parts of italy