Examples

Dataset consists of 100,000 papers from arXiv.org and the feature vector of each abstract text is reduced to 2D using PCA and UMAP. The data is clustered using HDBSCAN technique. Dataset size is 100 MB, may load slow.

Explore around 7,000 articles from The New York Times published between January 2022 and April 2022. Each article's text embedding is reduced to two dimensions using UMAP algorithms. The data is then reduced to three dimensions using PCA and clustered using HDBSCAN method. You can find this dataset on Kaggle.

Silk Road is the first modern darknet market and stands out as one of the most iconic cases in blockchain history. Try to identify anomalies in Silk Road crypto transactions with Cosmograph.

Generated graph sample (10,000 nodes, 19,800 edges).

Generated graph sample (100,000 nodes, 198,900 edges).