> ## Documentation Index
> Fetch the complete documentation index at: https://docs.cloudthinker.io/llms.txt
> Use this file to discover all available pages before exploring further.
# Code Review Analytics
> Explore PR volume, quality scores, finding distributions, and per-developer performance across all reviewed repositories
The Analytics section gives your team a data layer on top of every code review Oliver has run — surfacing where quality is improving, where findings cluster, and how individual contributors are trending over time.
Access it from **Code Review → Analytics** in the navigation.
## Analytics tabs
The Analytics section contains three tabs. The Leaderboard and Skill Matrix tabs are fully documented on their own reference pages; this page covers the Overview and Author Performance views.
| Tab | What it shows |
| ---------------- | ---------------------------------------------------------------------------------------------------------------------------------------- |
| **Overview** | Period-over-period PR volume, average Quality Score, finding volume by severity, and top finding categories |
| **Leaderboard** | Developer scores ranked by quality and code complexity — see [Leaderboard](/guide/code-review/leaderboard) |
| **Skill Matrix** | Domain proficiency heatmap across Security, Performance, Correctness, and Patterns — see [Skill Matrix](/guide/code-review/skill-matrix) |
## Overview tab
The Overview tab answers: *Is our code quality improving? Which finding types appear most often?*
### Summary cards
Four metric cards appear at the top of the Overview tab:
| Card | Description |
| ----------------- | ------------------------------------------------------------------------------------------------------------------ |
| PRs reviewed | Total pull requests that received a review in the selected period |
| Avg Quality Score | Mean quality score (1–10 scale) across all reviewed PRs; the score starts at 10 and is reduced by finding severity |
| Total findings | Sum of all findings, broken down by Critical, High, Medium, and Low severity |
| Resolution rate | Percentage of findings resolved before the PR was merged |
### Period comparison
Below the summary cards, weekly and monthly panels show current-period values alongside prior-period values with directional indicators (↑ / ↓). Read the combination of Quality Score and finding count together:
| Quality Score | Finding count | Interpretation |
| ------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------- |
| ↑ Rising | ↓ Falling | Developers are catching and fixing issues earlier |
| ↑ Rising | → Stable | Team is resolving a larger share of findings before merge |
| ↓ Falling | ↑ Rising | New contributors, new repositories, or reduced resolution effort — check [convention rules](/guide/code-review/convention-rules) coverage |
| → Stable | → Stable | Review quality is holding steady — use the Leaderboard to check for uneven distribution across developers |
### Finding trend chart
The trend chart plots finding volume by severity (Critical, High, Medium, Low) over the selected date range. Each severity renders as a separate series.
| Pattern | What it signals |
| ------------------------------------------- | ------------------------------------------------------------------------------------------------------------ |
| Critical and High declining over weeks | Developers are addressing severe issues before merge |
| Medium and Low rising while High stays flat | Higher PR volume or new convention rules are surfacing more style issues — expected when you first add rules |
| Sudden spike across all severities | A large refactor landed, a new repository was added, or a new contributor joined the review pool |
| Flat across all severities | Coverage and quality are stable; widen the date range to spot longer trends |
### Top finding categories
Below the trend chart, a ranked list shows which finding categories appear most frequently — for example, "Missing error handling", "Hardcoded secrets", or "N+1 queries".
The top categories are the highest-yield targets for convention rules. A single convention file targeting a pattern that appears in 40% of your findings eliminates that noise from future reviews. See [convention rules](/guide/code-review/convention-rules) for how to encode these patterns.
## Author performance tab
The Author Performance tab aggregates data across all reviewed pull requests, one row per developer. Use it alongside the Leaderboard for both relative team rankings and individual deep-dives.
### Author table columns
| Column | Description |
| ----------------- | ---------------------------------------------------------------------------------- |
| PRs reviewed | Total PRs that received an AI review for this author in the selected period |
| Avg Quality Score | Mean quality score (1–10) across all of this author's reviewed PRs |
| Total findings | Sum of findings by severity for this author |
| Resolution rate | Percentage of this author's findings resolved before merge |
| Impact score | Cumulative code complexity contributed — see the Leaderboard for the exact formula |
### Developer profile dialog
Click any row to open the full profile. Profile data uses a rolling 90-day window regardless of the tab's selected date range.
| Section | Description |
| --------------------- | -------------------------------------------------------------------------------------------------- |
| Profile header | Avatar, total MR count, total findings, average quality score, and overall Skill Matrix percentage |
| Skill Radar | Polar chart of the four Skill Matrix domains for this developer |
| Resolution Rate | Findings resolved before merge, displayed as a progress bar |
| Focus Areas | Auto-derived **Strengths** (top 2 domains) and **Development areas** (bottom 2 domains) |
| Weekly Findings Trend | Stacked column chart of findings by severity over the past 90 days |
| Pattern Force Graph | AI-identified recurring finding patterns visualized as a force-directed network |
Use author profiles alongside the [Leaderboard](/guide/code-review/leaderboard) to get both relative team rankings and individual deep-dives.
## Filters
All charts and tables respond to the filter bar at the top of the Analytics page:
| Filter | Effect |
| ---------- | ---------------------------------------------------------------------- |
| Date range | Standard presets (7d, 30d, 90d) or a custom range |
| Repository | Limit data to one or more connected repositories |
| Severity | Surface only Critical and High to focus on the most impactful findings |
## Related
Scoring formula that balances code quality with productivity across your team
Domain-level proficiency heatmap across Security, Performance, Correctness, and Patterns
Encode team standards to reduce recurring findings in future reviews