Research Atlases
Three interactive atlases showing patterns in research and public discussion. Each combines a map with views of the data, individual paper or author details, and comparisons between analytical methods.
JSTOR: AI ethics literature
A catalogue of 7,076 records mapped by similarities in their titles, abstracts, or combined abstract and full-text data. The atlas shows BERTopic and LDA topics, publication metadata, controlled keywords, coauthorship and correlation matrices.
Each view reports its data coverage. The text source determines the papers included and the embedding method used.
Twitter: discussions of AI
A network map of 164,291 matched Twitter authors, with eight- and twelve-topic models, communities, topic membership, sentiment and emotion. You can also explore follower and post counts, retweets sent and retweets received, with author search and 2D and 3D views.
Positions show network structure. The 3D view covers the largest connected component. Community labels describe prevalent post content.
PhilPapers: philosophy literature
A map of 69,400 English-language philosophy papers, with 2D and 3D views. It shows density-based clusters, k-means groups and LDA topics, alongside paper and author search and lists of semantically similar papers.
Nearby papers have similar text representations. Each clustering method groups papers according to its criteria, while their positions in the selected map stay fixed.