A random sample of 10,000 Anki collections, each containing over 5,000 review log entries, was compiled by open-spaced-repetition for the FSRS project. The dataset includes three parts: revlogs, cards, and decks, and was last updated on August 22, 2025. It likely contains a mix of older, active users and newer users.
Use Cases
- Modeling memory retention curves based on review history and time intervals.
- Analyzing user engagement and study patterns across different decks and cards.
- Training or evaluating spaced repetition scheduling algorithms using real-world review logs.
- Studying the relationship between card difficulty and review outcomes.
Strengths
- Contains over 5,000 review log entries per sampled collection, providing substantial individual user histories.
- Sampled from 10,000 collections, offering a broad base of user data.
- Explicitly designed to include a mix of older, active users and newer users.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count, file formats, and license information are unknown, which may limit suitability assessment.
- Data may reflect user bias inherent to the Anki platform and its user base.
Provenance
- Source
- Anki, a popular flashcard application.
- Collection Method
- Random sample of user collections meeting a minimum review log entry threshold.
- Time Range
- null
- Freshness
- Last updated 2025-08-22 06:42:25.
- Geography
- null