VoteCasters: Votes are cast.

• Parent Page: Election

VoteCasters

• Voter geometries cast their votes.
• A density of a population of voters is summed. A 1D or 2D spatial model is used to make the decision of how to vote or who to vote for.
• The vote is aggregated into a set that contains the information needed to elect, explain, and visualize.
• There are two ways to cast.
  1. By points.
  2. By regions. This is more complicated.
• The “votes” output data structure is described below. You can skip to that part.

Preferences

A single Preference

: also called a single vote.

These are the three ways to state a preference.

• Plurality: a choice of a candidate.
• Scores: a score for each candidate.
• Ranking: a rank for each candidate.

A preference is stored as data in the following ways:

• Plurality
  • pluralityVote: the chosen candidate.
    • A candidate is represented as an integer from 0 to a max. The max is the number of candidates minus 1.
    • type: Number
• Scores
  • scoreVote: an array of scores, indexed by candidate.
    • A score is a number between 0 and 1. In the future, maybe integers between 0 and a maximum score.
    • type: Number[], Number array
    • length: number of candidates
• Ranking
  • ranking: an array of ranks, indexed by candidate.
    • A rank of 0 is the best ranking, then 1, 2, and so on. No rankings are skipped, but skipping is allowed. Equal ranks are allowed.
    • type: Number[], Number array
    • length: number of candidates
  • cansByRank: an array of candidates lists, indexed by rank. Candidates can share a rank. Number[][]..
  • indexInOrder: an array of candidates in order of rank. Only one candidate is listed per rank. Number[].
  • netWinsPairwise: The number of wins minus number of losses for the first of a pair of candidates. Number[][]
  • bordaScores: The borda scores for each candidate for one vote. Number[]
  • bordaFractions: Fractional borda scores, between 0 and 1. Number[]

Preference Lists

• Plurality
  • There is no “pluralityVotes” list of preferences because it isn’t needed.
• Scores
  • scoreVotes: a list of scoreVote preferences, Number[][]. Each preference has a score for each candidate.
• Ranking
  • rankings: a list of preferences. Each preference has a rank for each candidate. Number[][]. 0 is the best ranking, then 1, 2, etc.
  • cansByRankList: a list of preferences. Each preference is a list indexed by rank. Each entry is a list of candidates at that ranking. Candidates are represented by index. Number[][][].

Counting

Ways to say how many votes.

• area: number of pixels.
• density: votes per pixel. From 0 to 1.
• voteCount: a number of votes.
• totalVotes: total number of votes.
• All of these are non-negative real numbers.

Preference Tallies

Preference tallies tell how many votes have a listed preference.

• voteCounts: a list of numbers of votes for a list of preferences.
  • Also used for “grid” of votes.
• voteFractions: a list of fractions for a list of preferences. voteCounts as a fraction of the total number of votes. From 0 to 1.

Candidate Tallies

Candidate tallies are just tallies indexed by candidate. This is a way to combine preferences. We have defined tallies for each type of preference:

• plurality
  • countByCan: The number of plurality votes for a candidate.
  • voteFractionByCan: The fraction of plurality votes for a candidate.
• score
  • scoreSumByCan: number of votes times score, summed for each candidate.
  • scoreAverageByCan: average score for each candidate. = scoreSumByCan / totalVotes. number of votes times score divided by max score, summed for each candidate. (not used)
  • scoreFractionAverageByCan: average fractional score for each candidate.
• ranking
  • bordaScoreSumByCan: sum of borda scores for a candidate.
  • bordaFractionSumByCan: The sum of fractional borda scores for a candidate. (not used)
  • bordaFractionAverageByCan: The average fractional borda score for a candidate.
  • firstPreferences: a sum of ranking votes for each candidate, counting the vote only by the first preference.

Outside of voteCasters, we sometimes use a loosely defined tally. It changes its definition depending on the type of vote.

• tallyFractions: A general term that should only be used when we don’t need to know what type of vote was being used but instead just want to give each candidate a number between 0 and 1.

• Uses of tallyFractions

  • tooltip uses tallyFractions and ranking
  • TestVoterGraphic uses tallyFractions
  • A general representation of a vote as a vector.

Pairwise Tallies

Pairwise tallies are indexed by a pair of candidates. The winning candidate is first and the losing candidate is second. This is another way to combine preferences.

• ranking
  • winsPairwise: The number of wins for the first of a pair of candidates.
  • winFractionPairwise: The fraction of wins for the first of a pair of candidates.

Votes

Some of the following object is output from a vote caster as the “votes” data structure.

• preferencesLists:
  • scoreVotes - Each preference has a score for each candidate.
  • rankings - Each preference has a rank for each candidate.
  • cansByRankList - Each preference is a list indexed by ranking. A list of candidates is at each ranking.
  • There is no “pluralityVotes” list of preferences because it isn’t needed. It would be an identity matrix.
• preferenceTallies:
  • voteFractions - the fraction of the voters who share the same preference.
• candidateTallies:
  • voteFractionsByCan - the fraction of plurality votes for each candidate.
  • scoreFractionAverageByCan - the average fractional score for each candidate.
  • firstPreferenceFractions - a list of fractions of voters who ranked a candidate first, indexed by candidate.
• pairwiseTallies:
  • winFractionPairwise - the fraction of wins for the first of a pair of candidates.
• votesByGeom:
  • Vote data for each voter geometry. A list of votesForGeom.
  • votesForGeom: Vote data for just one voter geometry. votesForGeom parts may include one of the following, depending on the vote caster:
    • votesForGeomCells: { ranking, cells }
    • votesForGeomIntervals: { ranking, intervalBorders }
    • votesForGeomGrid: { grid, voteSet, voterGeom }
      • This varies as well, depending on voteCaster.
• parties: { partiesByCan, numParties }
• numCans: the number of candidates