Visualising Culture at Scale: 2D Embeddings from 4 DR Algorithms

This repository contains 2D embeddings generated from four dimensionality reduction (DR) algorithms applied to four distinct cultural and humanities collections. Created by Giacomo Alliata for the Journal of Cultural Analytics in 2026, the data serves as a replication set for evaluating visualization techniques at scale. The original high-dimensional source data is withheld due to copyright restrictions, leaving only the coordinate outputs for analysis.

Use Cases

• Benchmarking the visual clustering performance of the provided 2D embeddings
• Comparing the spatial distribution patterns produced by the four dimensionality reduction algorithms
• Replicating the qualitative visualization evaluations described in the associated Journal of Cultural Analytics paper

Strengths

• Includes outputs from four different dimensionality reduction algorithms for direct comparison
• Provides replication data for specific cultural analytics research
• Covers four distinct, though anonymized, cultural collections

Limitations

• Original high-dimensional features are unavailable, preventing re-running the DR process
• Lack of descriptive metadata for the underlying cultural artifacts
• Restricted to 2D visualization outputs only

Provenance

Source

Journal of Cultural Analytics / Giacomo Alliata

Collection Method

Algorithmic generation via dimensionality reduction applied to private cultural collections