Generate Agile Metrics Dashboards

Agile Metrics Dashboard Agent designs and implements Agile metrics dashboards for Scrum teams, product managers, and stakeholders. It covers three metric categories - velocity and performance (story point velocity, performance vs. commitment, velocity coefficient of variation), flow and cycle time (lead time, cycle time, work in progress, throughput), and quality and technical health (defect density, escaped defects, technical debt ratio, code coverage) - and turns them into dashboard code, predictive models, and role-based views.

Use it when building or improving a Scrum team's metrics dashboard: collecting sprint data from Jira and Git, visualizing burndown, velocity, and cycle time, predicting next-sprint velocity statistically, or designing separate dashboard views for a scrum master, product owner, and development team. It is not a project-management tool itself - it produces the data-collection, visualization, and configuration code around metrics that live in Jira, Azure DevOps, and Git, not a replacement for those systems.

Input is historical sprint data (story points, transitions, defects) from a tracker and a repository, plus the stakeholder a dashboard is being built for. Output includes a JavaScript metrics collector class that pulls sprint and story data via Jira and Git APIs and computes velocity, burndown, cycle time, and defect rate; a React dashboard component (using recharts) with a sprint-health score, burndown chart, velocity trend, and cycle-time distribution by percentile; a Python velocity predictor using pandas, numpy, scipy, and scikit-learn's linear regression to compute the coefficient of variation, a confidence interval, and a trend-adjusted forecast for the next sprint's velocity; and a YAML dashboard configuration defining refresh intervals, widgets, and alert thresholds separately for scrum master, product owner, and development-team views.

Connects to JIRA/Azure DevOps APIs for story updates, Git for code-level metrics, CI/CD pipeline data for deployment frequency, and customer feedback metrics; the Python model uses pandas, numpy, scipy, and scikit-learn, and the dashboard component uses the recharts charting library:

    def calculate_velocity_statistics(self):
        velocities = self.data['velocity']
        return {
            'mean': velocities.mean(),
            'std': velocities.std(),
            'coefficient_of_variation': velocities.std() / velocities.mean(),
            'confidence_interval': stats.t.interval(
                0.80, len(velocities)-1, 
                loc=velocities.mean(), 
                scale=stats.sem(velocities)
            )
        }

Scrum masters, product owners, and engineering leads who need a working dashboard, not just a metric list - covering data-quality and validation rules, performance-optimized queries with caching and indexing, red-flag and green-indicator thresholds (for example a velocity drop over 20% or a cycle-time increase over 30%), and a governance process for keeping metric definitions and dashboard relevance current.