Sparse Matrix

• Polis gathers a sparse matrix of votes, as not every participant will vote on every comment.
• Thus, 👾 Algorithms used to analyze the data must be robust to sparsity.
• In practice, here's an entire conversation, from "Bowling Green Daily News Bowling Green, Kentucky, USA", found in the ⚗️ Case studies. You can see every single vote.
  • Each row is a participant's voting record
  • Each column is a comment
  • Blue represents an agree, red a disagree, and yellow a pass or 'did not see'.
  • Many participants in Polis conversations will not see many comments. This does not prevent dimensionality reduction and clustering 👾 Algorithms from making meaning by finding patterns