Public dataset
Danielle Strachman's public X cognitive graph
Every one of the 13,357 recovered public records is present. graph.json is the direct graph payload. tweets.ndjson is the streaming-friendly raw corpus. Start with manifest.json for counts and checksums.
13,357Tweets
13,409Nodes
16,076Edges
32Topics
| File | Bytes | Contents |
|---|---|---|
| /graph.json | 15,266,472 | Complete node and edge graph, including every recovered tweet as a node |
| /tweets.json | 13,680,391 | All recovered tweet rows joined to their assignments as one JSON array |
| /tweets.ndjson | 9,287,656 | All recovered tweet rows in streaming-friendly NDJSON |
| /assignments.ndjson | 3,187,777 | One deterministic assignment row per tweet |
| /model.json | 181,212 | Topic model metadata, 32 topics, and 12 interaction modes |
| /atlas.json | 62,170 | Seven-system synthesis and report evidence ledger |
Browser fetch is enabled cross-origin. Example const graph = await fetch('/graph.json').then(r => r.json()).