日本のTwitter界隈の規模と重複を可視化
| # | 界隈 | メンバー | アクティブ | 総リーチ | 中央値FL | インフルエンサー |
|---|
Jaccard類似度(0=重複なし、1=完全一致)
| 界隈A | 界隈B | 共通 | Jaccard | A→B | B→A |
|---|
界隈間のフォロー関係から算出した親和度。メンバーが他界隈のメンバーをどれだけフォローしているかで測定。
| 界隈A | 界隈B | A→B | B→A | 親和度 |
|---|
フォローベース界隈間親和度マトリクス
フォロー親和度から抽出した界隈クラスタ。弱い結線はノイズとして除外し、孤立界隈は単独表示しています。
Accounts seen across multiple communities or clusters. Useful for finding expansion hubs.
| Account | Type | Communities | Clusters | Followers | Score |
|---|
| Account | Type | Communities | Clusters | Followers | Score |
|---|
| Cluster A | Cluster B | Representative Accounts | Count |
|---|
Same ranking after removing nanpa as a source community.
| Account | Type | Communities | Clusters | Followers | Score |
|---|
Accounts shared across multiple current communities but not yet in the high-confidence member set.
| Account | Type | Communities | Clusters | Followers | Score |
|---|
More specific seeds: shared by 3-5 non-nanpa communities, 2-3 clusters, and at least 4 follow edges.
| Account | Type | Communities | Clusters | Followers | Score |
|---|
Filtered to accounts classified as community seeds instead of media-like generic hubs.
| Account | Type | Communities | Clusters | Followers | Score |
|---|
Media, official, or celebrity-style hubs kept separate so they do not dominate seed review.
| Account | Type | Communities | Clusters | Followers | Score |
|---|
Adjusted ranking with direct composite, second-order parent-overlap, and family coverage penalties.
| Proposal | Communities | Example Seeds | Actionable | Penalty | Novelty |
|---|
Proposals that exclude current composite communities such as promoted hybrid definitions.
| Proposal | Communities | Example Seeds | Actionable | Novelty |
|---|
Parent-community pairs already covered by promoted hybrid definitions.
| Family Pair | Covered By | Coverage |
|---|
界隈間のフォロー親和度をネットワークで可視化。ノードサイズ=メンバー数、エッジの太さ=親和度。