Home: README
- Learners ready to prove end-to-end capability.
- Engineers preparing portfolio-quality internal tools.
- Four graded capstones that map to real operational workflows.
- Completion of phases 0-6.
- Quality tooling baseline and working template.
Business goal:
- Standardize and merge operational spreadsheets safely.
Inputs/outputs:
- Input folder of
.xlsxfiles. - Outputs: master CSV/XLSX, rejects CSV, run log.
Implementation milestones:
- schema contract,
- normalization,
- validation,
- output writing,
- tests and logging.
Test scenarios:
- missing required columns,
- mixed header styles,
- malformed workbook.
Failure modes:
- hard crash on one bad file,
- silent row drops,
- overwritten outputs.
Pass/fail rubric:
- pass = deterministic outputs, rejects with reasons, logs, tests.
Stretch goals:
- config-driven schemas,
- richer Excel formatting.
Business goal:
- Move validated report data into durable SQL pipelines.
Inputs/outputs:
- Inputs from Capstone A.
- Outputs: staging load, reporting load, daily summary export.
Implementation milestones:
- table design,
- db connectivity,
- idempotent loading,
- summary query export.
Test scenarios:
- duplicate source data reruns,
- timeout and retry handling,
- schema mismatch.
Failure modes:
- duplicates in reporting table,
- partial writes without rollback,
- hidden auth assumptions.
Pass/fail rubric:
- pass = idempotent ETL, stable logging, recoverable failures.
Stretch goals:
- SQLAlchemy migration path,
- performance tuning and indexing.
Business goal:
- Ingest monitoring health data into reporting cache.
Inputs/outputs:
- Monitoring API reads.
- Outputs: cache tables + daily ops reports (xlsx/html).
Implementation milestones:
- read-only endpoint integration,
- field mapping worksheet,
- cache table writes,
- freshness checks.
Test scenarios:
- API auth failures,
- endpoint timeout,
- payload field drift.
Failure modes:
- stale data with no warning,
- mixed source semantics,
- over-polling source systems.
Pass/fail rubric:
- pass = stable read ingestion, clear ownership mapping, stale-data signaling.
Stretch goals:
- Teams/email summary delivery,
- adaptive polling windows.
Business goal:
- Deliver usable operational visibility to non-technical users.
Inputs/outputs:
- Database summary/cache data.
- Output: browser UI with filters and exports.
Implementation milestones:
- user story definition,
- baseline UI,
- SQL-backed filters,
- exports and freshness indicators,
- deployment notes.
Test scenarios:
- empty data windows,
- stale cache,
- high row counts.
Failure modes:
- slow loads,
- unclear filter behavior,
- no ownership/hand-off docs.
Pass/fail rubric:
- pass = user can answer core daily questions without SQL access.
Stretch goals:
- role-based views,
- API service layer.
- A practical portfolio proving SME-level operational Python capability.
- Force one capstone dependency failure and recover without data corruption.
- Simulate stale data and communicate impact clearly.
- Rerun all capstones with same input and prove determinism.
- project sprawl:
- isolate each capstone boundary and dependencies.
- weak test confidence:
- add scenario-based tests for failure paths.
- handoff risk:
- strengthen runbooks and ownership metadata.
You are SME-ready when you can:
- demo all capstones end-to-end,
- explain architecture tradeoffs,
- support and recover production-like failures.
- Play: add one meaningful enhancement per capstone.
- Build: complete base rubric before enhancements.
- Dissect: produce architecture diagrams and data flow notes.
- Teach-back: run a capstone walkthrough for peer review.