Transport operations background • SQL • Python • dbt • Dashboards • Data quality
I am a logistics engineer and former Transport Fleet Coordinator building toward junior roles in supply chain analytics, operations analysis, data analysis, and BI. My formal logistics experience provides the domain context; the repositories below are independent, executable work samples rather than corporate or production work.
| A manager may ask | Direct evidence | Transferable capability |
|---|---|---|
| KPI's | Control Tower issue #6 → tested PR #7 → manager case study | Metric governance, SQL/Python debugging, and root-cause validation |
| Turn operational records into a usable report? | Logistics issue #3 → export PR #4 → verified screenshot | KPI design, filtering, visualization, and operational communication |
| build trustworthy reporting models | dbt issue #3 → evidence PR #4 → validation summary | Dimensional modeling, reusable SQL, tests, documentation, and lineage |
| receive a change request and prove the result | The linked issues and PRs document the problem, acceptance criteria, implementation, tests, and limitations | Ticket-based delivery, regression testing, and clear handoff |
Business problem: service, fulfillment, inventory, transportation, and planning metrics can disagree when their grain or definitions are unclear.
Delivered: a one-command local pipeline over 2,400 synthetic orders and 5,928 order lines, with independent pandas and DuckDB implementations.
Evidence: eight reconciled SQL/Python metrics, 38 automated data-quality checks, eleven regression tests, exception outputs, and a bounded executive summary.
Transferable to: KPI reconciliation, recurring operational reporting, carrier or supplier investigation, and data-quality troubleshooting.
Business problem: operations teams need service and cost measures that respond consistently to the same period, supplier, and route selections.
Delivered: a Streamlit/Plotly application calculated from 996 deterministic synthetic orders, with five governed KPIs and operational observations.
Evidence: verified dashboard image, known-result KPI fixtures, input contracts, sixteen tests, application smoke coverage, and GitHub Actions.
Transferable to: operational dashboards, recurring KPI reporting, exception prioritization, and communication with non-technical stakeholders.
Business problem: dashboards need consistent, documented models between raw operational sources and business-facing metrics.
Delivered: deterministic source generation plus staging, intermediate, fact, dimension, operations, and customer models on dbt and DuckDB.
Evidence: 20 models, a Type 2 snapshot, two exposures, 71 passing data tests, 98.35% unit-weighted fill reconciliation, generated documentation, and CI evidence.
Transferable to: analytics model maintenance, SQL transformation, lineage, regression testing, and reliable BI datasets.
| Work activity | Demonstrated approach |
|---|---|
| Define a KPI | Document formula, grain, denominator, unit, and limitations |
| Investigate an exception | Trace source rows through transformations to the reported result |
| Change a business rule | Update both implementations, tests, documentation, and reconciliation |
| Prepare an operational review | Separate observations, possible actions, assumptions, and missing context |
| Maintain recurring reporting | Use deterministic runs, validation contracts, regression tests, and CI |
- Supply chain: OTIF, on-time delivery, fill rate, inventory, transportation, forecast, service, and cost metrics.
- SQL and modeling: joins, CTEs, aggregations, window functions, dimensional models, marts, and reconciliation queries.
- Python: pandas analysis, validation, KPI calculation, reporting, pytest, Streamlit, and Plotly.
- Analytics engineering: dbt models, snapshots, tests, documentation, lineage, exposures, DuckDB, Git, and GitHub Actions CI.
The featured datasets are synthetic and the projects run locally. They demonstrate junior-level implementation and explainable work samples, not enterprise production experience, real customer impact, Power BI/Tableau delivery, SAP integration, cloud operations, or years of professional data-platform ownership.
AI tools assisted implementation and review. I remain responsible for the problem definition, metric choices, validation, testing, corrections, and explanation of the final work, and I can reproduce and modify the featured projects.