Filters and views

Filtering your data

The sidebar on the left side of your dashboard contains all the filters. Every chart and metric on the page updates in real time as you change your selections. Filters apply across both the overview and the commit explorer.

Date range

The date range controls which commits are included in the analysis. Four presets are available: 7 days, 30 days, 90 days, and 1 year. The default is 30 days.

The selected range is shown at the top of the date section as a display with both the start/end dates and the number of calendar days and working days in the range. Working days are used as the denominator for density and active days calculations.

The pre-AI baseline always uses the same window length. If you select a 30-day range, the baseline comparison uses the 30-day period immediately before your baseline date. This keeps comparisons fair regardless of which range you're viewing.

Outlier filter

The outlier filter removes commits at the extreme end of the line-change distribution. Two options are available: P95 (hide the top 5% of commits by size) and P99 (hide the top 1%).

This filter is particularly useful for teams whose repositories occasionally receive large automated commits, library upgrades, or generated code dumps that would otherwise distort the line change charts and churn figures. The filter applies only to chart rendering and aggregate stats; the commit explorer always shows all commits regardless of the outlier filter setting.

The P99 filter is the default. Most teams find it appropriate for removing genuine outliers without hiding meaningful large commits. Use P95 if your codebase regularly includes large but legitimate human-authored commits that you want to exclude from trend analysis.

Repositories

All connected repositories are listed here. Each shows a commit count for the current date range. Toggle individual repositories on or off to focus the dashboard on a specific subset of your codebase.

The All and None buttons at the top of the list let you quickly select or deselect everything, which is useful when you want to isolate a single repository from a large set.

Branches

Branches are listed in order of commit volume within the selected date range. By default, all branches are visible. Most teams find it useful to filter to main and develop (or your equivalent primary branches) to get a signal that reflects work that has actually been reviewed and merged.

Feature branches can show elevated churn and burst relative to primary branches: this is normal, since feature branches are where exploratory work happens. Including them in your analysis is fine, but be aware of the effect on aggregate metrics.

Authors

Each connected author is listed with their commit count for the current date range. Authors are shown with their colour swatch (the same colour used to identify them in all charts).

Filtering to a subset of authors is useful for comparing specific contributors, or for excluding bot accounts that couldn't be automatically excluded during import. If you have a CI/CD bot that commits frequently, excluding it here will give you a cleaner view of human-authored activity.

How filters interact

All filters apply simultaneously. Selecting a single repository and a single author shows only that author's commits to that repository in the selected date range. The commit count shown next to each filter item reflects the current selection, so as you change other filters the counts update to reflect only matching commits.

Filter state is preserved when you navigate from the overview to an author deep-dive and back. It is not preserved between sessions: each new session starts with all filters set to their defaults.