Views and unique views
Views are repository page requests reported through GitHub's traffic data. Unique views are GitHub's aggregate estimate of distinct viewers within the reported interval. The API does not return account identities behind those counts.
A person can generate several views, and a daily unique estimate cannot safely be added across days to create a monthly person count. README images served from another origin are also not identifiable as GitHub traffic viewers.
Clones and unique clones
Clones count full repository clone events reported by GitHub. Routine fetches are not included. Unique clones are GitHub's aggregate estimate of distinct cloners within the interval.
A developer may clone more than once, and automated systems can clone repositories for CI, packaging, mirroring, security scanning, or deployment. A clone therefore signals repository acquisition, not necessarily a new human evaluator or product installation.
Patterns worth investigating
Views rising faster than clones can indicate that documentation, issues, release notes, or a linked path is attracting readers. Clones rising with flat views may point to automation or a workflow that reaches the repository without repeated browsing. Both rising together can justify a closer look at referrers, popular paths, stars, and release timing.
These are hypotheses, not diagnoses. Compare matching date ranges, preserve UTC boundaries, and look for repeated patterns across more than one interval.
- High views, low clones: inspect popular content and documentation
- High clones, low views: investigate automation and distribution workflows
- High unique ratios: remember they remain upstream estimates
- Sudden spikes: compare referrers and known project events
Use exact language in reports
Call the metrics views, unique views, clones, and unique clones. Avoid relabeling unique estimates as people, customers, conversions, or installations. State the source, generation time, date range, and any missing observations.
RepoLogbook preserves these repository-level aggregates and provides charts plus an exact-value table. Its methodology does not infer visitor identity or causal attribution.
Frequently asked questions
What counts as a GitHub repository view?
A view is a repository page view reported by GitHub. It is not the same as a README image request and does not reveal the viewer's identity.
What counts as a clone?
GitHub's traffic documentation describes full repository clones and explicitly excludes ordinary fetches.
Why can clones exceed unique cloners?
One aggregate unique cloner may clone more than once during the interval, so total clone events can exceed the unique estimate.
Do views or clones prove product adoption?
No. They are attention and acquisition signals. Automation, CI, mirrors, repeated activity, and incomplete context can affect interpretation.