Measuring crypto narrative signals from X (Twitter)
We score public posts with a dual-model sentiment pipeline (BERTweet + DistilBERT), aggregate activity and silence into time windows, and compare the results with market factors. Findings are reported as measured-positive or negative.
ENTRIES
751,500
DAYS ARCHIVED
298
LAST UPDATE
Oct 8
Research
Methodology and collection framework
Ethical data sourcing, analysis approach, and documentation standards
Dataset
Structure and access information
Schema, formats, licensing terms, and distribution details
Access
Subscribe to dataset tiers
Choose your tier and get daily crypto sentiment data via Patreon