Israeli society has experienced significant political polarization in recent years, reflected in five Knesset elections held within a four-year period (2019-2022). Public discourse increasingly references hypothetical divisions of the country into politically homogeneous "cantons." This paper develops a data-driven algorithmic approach to explore such divisions using publicly available municipality-level election results and geographic boundary data from the Israel Central Bureau of Statistics. We partition 229 Israeli municipalities into geographically contiguous cantons that maximize internal political similarity. Our methodology employs four clustering algorithms -- Simulated Annealing, Agglomerative Clustering with contiguity constraints, Louvain Community Detection, and K-Means (baseline) -- evaluated across four feature representations (BlocShares, RawParty, PCA, NMF), three distance metrics (Euclidean, Cosine, Jensen-Shannon), and six values of K (3-20), yielding 264 experimental configurations. Key results show that BlocShares with Euclidean distance and Agglomerative clustering produces the highest clustering quality (silhouette score 0.905), while NMF with Louvain community detection achieves the best balance between political homogeneity, silhouette quality (0.121), and interpretable canton assignments. Temporal stability analysis across all five elections reveals that deterministic algorithms produce near-perfectly stable partitions (ARI up to 1.0), while Israel's political geography remains structurally consistent despite electoral volatility. The resulting K=5 partition identifies five politically coherent regions -- a center-leaning metropolitan core, a right-wing southern arc, a right-leaning northern mixed region, and two Arab-majority cantons -- closely reflecting known political-demographic divisions. An interactive web application accompanies this work.

翻译：近年来，以色列社会经历了显著的政治极化，这体现在四年期间（2019-2022年）举行的五次议会选举中。公共讨论越来越多地提及将国家划分为政治同质化的“行政区”这一假设性方案。本文开发了一种数据驱动的算法方法，利用以色列中央统计局公开的市级选举结果和地理边界数据来探索此类划分方案。我们将229个以色列城市划分为地理上连续且内部政治相似性最大化的行政区。我们的方法采用了四种聚类算法——模拟退火、具有邻接约束的凝聚聚类、Louvain社区检测和K-Means（基线）——在四种特征表示（BlocShares、RawParty、PCA、NMF）、三种距离度量（欧几里得距离、余弦距离、Jensen-Shannon散度）和六个K值（3-20）下进行评估，共产生264种实验配置。关键结果表明，采用欧几里得距离的BlocShares特征与凝聚聚类相结合能产生最高的聚类质量（轮廓系数0.905），而NMF特征与Louvain社区检测则在政治同质性、轮廓系数质量（0.121）和可解释的行政区划分之间达到了最佳平衡。对所有五次选举的时间稳定性分析表明，确定性算法能产生近乎完全稳定的划分（调整兰德指数最高达1.0），尽管选举存在波动，以色列的政治地理结构仍保持一致性。最终得到的K=5划分识别出五个政治连贯的区域——一个中间派倾向的大都市核心区、一个右翼的南部弧带、一个右倾的北部混合区以及两个阿拉伯人口占多数的行政区——这密切反映了已知的政治-人口分布格局。本研究还附带了一个交互式网络应用程序。