Emily18 Com Full Sets -2021-

| Cluster ID | Dominant Modality | Size (items) | Representative Themes (LDA keywords) | |------------|-------------------|--------------|---------------------------------------| | C1 | Text‑heavy (70 % transcripts) | 322 | “memory”, “family”, “childhood”, “storytelling”, “nostalgia” | | C2 | Image‑centric | 254 | “landscape”, “architecture”, “light”, “color”, “composition” | | C3 | Audio‑rich (58 % MP3) | 210 | “interview”, “soundscape”, “ambient”, “dialogue”, “field‑recording” | | C4 (Noise) | Mixed | 12 | — |

Visual inspection of the UMAP plots shows clear separation between C1–C3, confirming that multimodal embeddings preserve thematic distinctions. Emily18 Com Full Sets -2021-

All items were downloaded from the official Emily18 Com repository (https://archive.emily18.com/2021/full‑sets) under a CC‑BY‑4.0 license. The repository provides a SHA‑256 checksum for each file; integrity was verified before ingestion. | Cluster ID | Dominant Modality | Size

The transition aligns with the collective’s publicly stated “seasonal focus” (see Emily18 blog post, 2021‑04‑02). integrity was verified before ingestion.