voting advice applications:
data:
algorithm:
- per motion calculate distance between parties
- in a loop:
- calculate mean fit per party based on filled motions
- calculate distance between parties based on calculated mean scores
- per motion calculate distance between parties for that motion / distance between them based on the user's calculated mean fit
- per motion sum the above figure into a score expressing how well the motion helps distinguish best-fitting parties for the user
- ask user to answer the unanswered motion with the highest score
- maybe figure out some kind of confidence level to figure out when to advice the user to settle?