Where is my risky / complex code?¶

Understand is deliberately metric-agnostic — it reports the numbers and lets your team decide what "risky" means. This page shows the fastest ways to surface your hotspots — the most complex, largest, and most violation-heavy entities — from a one-glance overview down to a sortable per-entity export.

Pick the tool to match the question

You want to…	Use
A one-screen overview of where risk concentrates	The Home dashboard
See big and complex code visually, project-wide	Metrics Treemap
Flag entities that cross risk thresholds	Metrics Heatmap
Rank every entity by a metric and sort/filter	Export Metrics → CSV
Colour a call/dependency graph by complexity	Style a graph

Start from the Home dashboard¶

Opening a project's Home page (Project menu → Home, the house icon) gives you ready-made "Most Complex Functions", "Most Complex Classes", "Most Complex Files", and "Largest …" cards, plus a graded Function Complexity summary. It's the quickest way to see whether complexity is a problem and where it clusters. See Reading the Home dashboard for what each card and grade means.

Metrics Treemap — big + complex at a glance¶

Open Metrics → Metrics Treemap. A treemap draws one box per entity where box size and box colour are each mapped to a metric you choose, so large, hot-coloured boxes are your big, complex units — the hotspots to refactor first.

In the treemap options dialog you control:

Map metrics for: — the entity kind to show (files, classes, functions).
Group by: — an architecture to nest the boxes by (e.g. Directory Structure).
Map Size to: — the metric that drives box area.
Map Color to: — the metric that drives box colour, with Min Color / Max Color and a Use logarithmic Scale option that pulls extreme outliers into view.

A metrics treemap of the whole project, grouped by directory. Each file is a box sized by lines of code; the darkest boxes are the most complex files. — The default Metrics Treemap: box size = lines of code, colour = max cyclomatic complexity, grouped by directory — the big dark boxes are the refactoring candidates.

Out of the box the treemap is already tuned for risk: colour defaults to a complexity metric (MaxCyclomatic for files/classes, Cyclomatic for functions) and size defaults to lines of code (CountLine). Change either combo to re-map the picture — e.g. colour by CountCCViol to spotlight violation-heavy files. See what Cyclomatic Complexity means before you read too much into the colours.

Changes Treemap

The Changes Treemap (on the Compare menu) maps the change in a metric between two versions of your project, so you can see where complexity grew (or shrank) between two points in time.

Metrics Heatmap — threshold-based risk flags¶

Open Metrics → Metrics Heatmap → By \<architecture> (or right-click an architecture or entity and choose Metrics Heatmap). The heatmap lays out entities against a set of metrics and colours each cell on a pass→fail scale, so entities that cross a risk threshold light up.

The default heatmap flags risk on metrics including MaxCyclomatic (pass 10 / fail 50), MaxNesting (pass 7 / fail 10), CountCCViol (CodeCheck violations, pass 0 / fail 100), CountAnalysisError, and comment-to-code ratio. High-complexity, high-violation entities show hot; clean ones stay cool — a hotspot report for the whole architecture at a glance.

Rank every entity: the Metrics Browser and CSV export¶

Metrics → Browse Metrics opens the Metrics Browser — a metric-name → value summary per architecture and per selected entity. It's great for reading aggregates (project totals, per-file averages), but it is not a sortable per-entity ranking.
To rank every entity by any metric, export and sort externally. From the Metrics Browser toolbar use Export to CSV / Export to HTML (or Metrics → Export Metrics). Then sort the CSV by, say, Cyclomatic descending to get your worst offenders first.
Headless, und metrics myProject.und writes the project's configured metrics, one row per entity, to CSV. Choose the columns with the metric settings — see Export & track metrics and Run Understand from the command line.

For other ways to rank entities by a metric — an in-GUI sortable column, or a per-kind chart/histogram graph — see Rank & compare entities by a metric.

Enforce a complexity limit¶

Once you know where the complexity is, you can make crossing a threshold fail a check rather than relying on eyeballs: CodeCheck ships with complexity-based and structure-based rules you can turn on and run in CI. See What is CodeCheck? and View & triage violations.