donnemartin/system-design-primer
Learn how to design large-scale systems. Prep for the system design interview. Includes Anki flashcards.
Experimentation Intensity Score
45.7 / 100
A single, comparable figure that combines engagement, regularity, issue refinement and integration into one view of a repository's development activity — calculated fresh relative to every other repository currently in the database, so scores stay meaningful as more repositories are added.
How is this calculated, and why did I get this score?
- Why you got this score
- This score of 45.7/100 is the average of 4 applicable dimensions: Engagement (0.02), Regularity (0.91), Issue Refinement (0.52), Integration (0.37).
- How it's calculated
- Each dimension is min–max normalised against every repository currently in the database (reverse-coded where a lower raw value is the stronger pattern), then the applicable dimensions are averaged and scaled to 0–100.
- What it represents
- One comparable figure for how much iterative, structured development activity a repository shows — not code quality or project success.
- How to interpret it
- It's relative to whatever repositories are currently stored, so treat it as a within-this-dataset ranking aid, not a universal benchmark.
- How to use it
- Compare repositories side by side, and check the four dimension scores and metric cards below for the evidence behind the number. Full explanation →
Commit activity
Feeds into: Engagement & Regularity
- Commit frequency
- 0.27 / week
- Active-day ratio
- 2.8%
- Commit interval variability
- 1.7 days
- Longest inactivity gap
- 4.2 days
Issue activity
Feeds into: Issue Refinement
- Total issues
- 31
- Issue closure rate
- 41.9%
- Mean comments per issue
- 0.48
- Mean issue resolution time
- 15.0 days
Pull request activity
Feeds into: Integration
- Total pull requests
- 58
- Merge rate
- 6.9%
- Reviewed PR ratio
- 25.9%
- Mean PR cycle time
- 11.2 days
Commit activity over time
Learning Quality Indicator
83.5%
Share of classified repository text — commit messages, issues, pull requests and reviews — that reads as substantive learning-oriented language (problem identification, experimentation, reflection or refinement) rather than routine text.
How is this calculated, and why did I get this score?
- Why you got this score
- Of the 121 pieces of repository text analysed, 101 (83.5%) were classified as substantive learning-oriented language, most often 'problem identification' (50); the remaining 20 were classified as routine text.
- How it's calculated
- Every classified text snippet is assigned one of five categories by a TF-IDF + classifier: Problem Identification, Experimentation, Reflection, Refinement, or None. The indicator is the percentage that fell into one of the four substantive categories.
- What it represents
- An estimate of how much repository text reflects learning-oriented language vs. routine/administrative text.
- How to interpret it
- The current classifier (Linear SVC) was trained on 100 labelled examples and measured 49.6% accuracy in cross-validation. Treat classifications as exploratory suggestions, not certainties.
- How to use it
- As a starting point for which artefacts might be worth reading directly — not proof that learning occurred. Full explanation →