A new paper on arXiv proposes a framework for quantifying gerrymandering using persistent homology, a tool from topological data analysis. The method extends the level-set filtration of Feng & Porter to compare voting data at both precinct and district levels. Because precincts are the smallest geographic units for which votes are reported, the approach aims to capture distortion that coarser district-level data might miss.
The authors demonstrate the framework on four states: California, Louisiana, North Carolina, and Texas. By examining how voting patterns change across scales, the method can flag irregularities that may indicate partisan boundary manipulation. The paper does not claim to prove intent, but it offers a quantitative signature of gerrymandering that could complement existing legal and statistical tests.
Since this is the only source in the digest, there is no second study to compare against. The work is presented as a methodological contribution, with the four states serving as case studies rather than a comprehensive national audit. Future work would likely need to validate the approach against known gerrymandering cases and other detection methods. As it stands, the paper adds a new geometric lens to a longstanding political problem.{