GitLab Analytics Platform
Executive Power BI dashboard backed by a custom Python ETL pulling cross-program engineering metrics from the GitLab API.
Overview
An executive Power BI dashboard at X-Energy that surfaces cross-program engineering metrics — issue throughput, MR cycle time, pipeline health, and velocity — across nearly every engineering group in the company. Backed by a custom Python ETL service I wrote that hits the GitLab API on a schedule, normalizes the data, and writes it to the analytics layer.
Architecture
- Ingest: Python ETL service that paginates the GitLab API across all configured groups and projects. Handles rate limits, partial failures, and incremental fetches.
- Normalize: Per-resource transforms (issues, MRs, pipelines, jobs) into a flat star-schema-friendly shape.
- Store: Persisted to a structured backing store consumed by Power BI.
- Surface: Power BI dashboard with executive-level rollups (program throughput, blockers, cross-team flows) and drill-down panes per team / project.
Why it matters
Before this, leadership had no consolidated view of engineering throughput across programs. Each group reported its own metrics in its own shape, on its own cadence. The dashboard surfaced patterns invisible to any single team — cross-program blockers, schedule slippage that only appeared when MR cycle time and pipeline failures were viewed together, and quantified velocity trends that fed into Earned Value reporting.
Stack
- ETL: Python, GitLab REST API, scheduled execution
- Storage: Structured analytics layer compatible with Power BI imports
- Dashboard: Power BI, DAX measures, drill-through reports
- Domain: Engineering throughput, MR cycle time, pipeline reliability, cross-program velocity
Internal to X-Energy. Architecture described here at the level of the resume.