What the charts are showing you
The overview page contains four charts. Each one surfaces a different dimension of your team's output. This page explains how to read each chart and what patterns are worth paying attention to.
Commits over time
The commits-over-time chart shows daily (or weekly) commit volume as stacked bars, with each author represented by their own colour. The height of the full bar on any given day shows total team commits. The individual colour segments show each contributor's share.
The horizontal dashed line running across the chart is your pre-AI daily average: the average number of commits per day before your baseline date. Days where the bar exceeds this line represent above-baseline output; days below it represent below-baseline output.
What to look for
A healthy pattern post-AI adoption typically shows bars consistently above the pre-AI average, with relatively even distribution across contributors. Uneven patterns (one author dominating the bars, or large spikes followed by flat periods) are worth investigating alongside churn data to understand whether the volume is translating to quality.
Use the Day/Week toggle in the top-right of the chart to smooth out noise. The weekly view is more useful for spotting trends; the daily view is better for identifying specific dates that look anomalous.
Line changes over time
This chart shows additions and deletions as a diverging view: additions extend upward from the centre line, deletions extend downward. Each bar represents a single day.
The centre line represents zero. A day with equal additions and deletions produces bars of equal height above and below the line. A day dominated by additions produces a taller upward bar. A day dominated by deletions (a cleanup or refactoring day) produces a taller downward bar.
What to look for
The overall shape tells you whether your team is in a building phase, a consolidation phase, or an unstable cycle. Consistent upward bars with moderate downward bars suggest healthy, progressive development. Large downward bars following large upward bars suggest code that was written and then substantially revised or removed, which connects to churn.
Outlier commits (large library additions, automated code generation) can dominate this chart. If a single day looks dramatically different from surrounding days, check whether the outlier filter is removing the most extreme commits.
Additions vs deletions by author
This chart shows each author as a pie, where the green segment represents additions and the red segment represents deletions, as a proportion of their total lines touched. The size of each pie is proportional to the author's total lines changed, so a larger pie indicates a higher-volume contributor in the selected period.
A faded smaller circle to the right of each pie shows the same author's profile during the pre-AI baseline period. This lets you see at a glance whether the ratio of additions to deletions has shifted since AI adoption.
What to look for
An author whose pie has grown significantly larger since the baseline period is shipping more volume. Whether that's positive depends on whether their churn ratio has moved in the same direction. An author with a much larger pie and a noticeably higher proportion of deletions may be generating code that frequently needs immediate correction.
Pies that are roughly the same size as their baseline shadow suggest stable output patterns regardless of AI adoption, which can mean the tools aren't being used, or that they're being used in a way that doesn't change visible output.
The pre-AI baseline line
The dashed line appearing in the commits-over-time chart is calculated from commits made before your baseline date. It represents the daily average across the pre-AI period, adjusted for the same date range you're currently viewing.
All percentage comparisons shown in the summary stats row (for example, "↑ 12% vs pre-AI") are calculated against this baseline. The baseline average is shown as a label at the right end of the dashed line.