Dashboard Overview

Your Central View

The GitVelocity dashboard is the first thing you see after signing in. It provides an organization-level summary of engineering velocity, giving managers and team leads a single place to understand what their teams are shipping and how much engineering ships over time.

The dashboard surfaces the leaderboard, recent activity, and recognition on one screen so leads don't have to context-switch to see what their team is shipping.

Time Range Filters

Use the time range selector to focus the dashboard on the period that matters most. Built-in options include 7-day, 30-day, and 90-day windows, along with a custom date picker for ad-hoc analysis. All charts, summaries, and breakdowns update instantly when you change the range.

Repository Breakdown

The repository breakdown shows velocity metrics for each connected repository side by side. Compare average PR scores, total velocity, and PR volume across repositories to identify where the heaviest engineering is happening and how velocity is distributed.

Team and Contributor Summary

Summary cards highlight top contributors, total PRs scored, and aggregate velocity for the selected time range. At a glance, you can see who is active, which teams are shipping, and how each person's velocity contributes to the whole.

Filtering by Language

A language selector sits beside the team performance chart. It narrows the trend to changes where one language made up the biggest part of the diff, so you can follow a single stack over time instead of the blended average of everything your team ships.

Further down, the Score by Language chart compares your busiest languages against each other, either as a monthly total or per change. Score tracks change size as much as difficulty, so config languages sit low because config diffs are small. Languages covers how the label is picked and how to read that chart.

Zooming in on Multiplier and Score by Language

Drag across either chart to zoom into a time range, then hold Shift and drag to pan across nearby periods. On a touchscreen, spread or pinch two fingers to zoom in or out, and drag horizontally with one finger to pan. Vertical swipes scroll the page. Fit adjusts the vertical scale to the values visible in that range. On Multiplier, it includes the 1× baseline and team percentile band; on Score by Language in Total mode, it fits the stacked totals from zero.

Fit and Reset appear once you zoom in and disappear when you return to the full view. This also applies to the Performance Trends charts. Reset restores that chart's full time range and default vertical scale. It keeps your metric, baseline settings, and selected languages. Changing the metric, organization, or time window resets the view automatically; hiding or adding a language keeps your zoom.

Score Distribution

A distribution chart groups scored PRs into buckets, showing how your team's work falls across the 0--100 scale. This makes it easy to spot patterns -- for example, whether the team is mostly shipping small incremental changes or tackling larger features.

Trend Lines

Trend lines track velocity over time so you can see whether velocity is increasing, plateauing, or shifting in character. Use trends to correlate velocity changes with team events like sprint boundaries, new hires, or priority shifts.

Reading a chart with a large team

Once a chart carries more than about a dozen people, color stops doing its job -- there are only so many line colors the eye can tell apart at a glance, and past that point every line looks like every other line.

So on a larger team the chart offers to highlight the contributors who shipped half the engineering across the range it draws, and fade the rest. It does not do this on its own -- every line starts at full strength, because deciding who matters before you have asked is not the chart's job.

The sentence above the chart says how many people that is: "5 key contributors shipped half the engineering over the last 12 weeks." The count in that sentence is the control -- select it and the charts show only those people in color. Select it again to bring everyone back.

When it is on, nobody is removed: every contributor is still drawn, still in the list beside the chart, and still counted in every total. The faded lines are the same data, shown quietly so the shape of the chart is readable.

The range matters. Highlighting is judged across the whole span on screen, not the newest bar alone -- so somebody who carried the team for two months and shipped nothing this week is still highlighted, and you can see where they stopped. Judging it on the newest bar would have faded their entire line, which is the one case you most want to notice.

Each chart decides this for its own range, which is why the weekly and quarterly views can highlight different people. A quarter can be concentrated in a few contributors while the last twelve weeks were spread evenly across the team -- that difference is worth seeing, not smoothing over.

Selecting specific contributors from the list takes over instead: your selection wins, and the charts highlight exactly who you picked.

When somebody stops

A small triangle on a line marks the last period before that person's velocity fell off and stayed down. It appears only where the fall is large enough to be worth a reader's attention and where that person was shipping enough to have been carrying real weight -- on most charts, no line carries one.

It is deliberately quiet about why. Someone who has stopped shipping, someone who left, and someone on leave all draw the same mark, and the chart has no way to tell them apart. It marks where to look, not what to conclude.

When the scoring model changed

A dashed amber line with a diamond at the top marks the date your PRs started being scored by a different AI model. That happens when the GitVelocity default moves to a new model, or when someone in your organization picks a different one in Settings → AI Scoring. Different models do not score on one scale, so a step in the lines at that date comes from the model, not from your team.

Hover over the diamond, tap it, or tab to it, to see which models changed and by roughly how much. When changes sit too close together to show a diamond each, one diamond covers them all and lists each change. The same line appears on every chart that plots scores over time: the Dashboard trends, Multiplier, Score by Language, contributor profiles and Benchmarks. On Multiplier, the vertical line with a diamond at the top is the model change; the vertical line with a label beside it is your baseline month. The model-change line disappears once your whole history is scored by one model; see Do I have to rescore my old PRs when GitVelocity switches to a new AI model?

Frequently Asked Questions

How often does the dashboard update?

Scores appear within a few minutes of a PR merging. The dashboard reflects real-time data -- there is no overnight batch process. As soon as the AI finishes scoring a PR (approximately 30 seconds), it shows up.

Can I filter by team or repository?

Yes. The dashboard supports filtering by repository, time range, contributor, and programming language. You can compare velocity across different repositories or zoom in on a specific team's velocity.