10 - Capstone Projects (Graded Specs for SME Readiness)¶
Home: README
Who this is for¶
- Learners ready to prove end-to-end capability.
- Engineers preparing portfolio-quality internal tools.
What you will build¶
- Four graded capstones that map to real operational workflows.
Prerequisites¶
- Completion of phases 0-6.
- Quality tooling baseline and working template.
Step-by-step lab pack¶
Capstone A - Excel Merger + Validator + Report¶
Business goal: - Standardize and merge operational spreadsheets safely.
Inputs/outputs:
- Input folder of .xlsx files.
- Outputs: master CSV/XLSX, rejects CSV, run log.
Implementation milestones: 1. schema contract, 2. normalization, 3. validation, 4. output writing, 5. 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.
Capstone B - ETL to SQL Database¶
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: 1. table design, 2. db connectivity, 3. idempotent loading, 4. 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.
Capstone C - Monitoring API Daily Ops Pipeline¶
Business goal: - Ingest monitoring health data into reporting cache.
Inputs/outputs: - Monitoring API reads. - Outputs: cache tables + daily ops reports (xlsx/html).
Implementation milestones: 1. read-only endpoint integration, 2. field mapping worksheet, 3. cache table writes, 4. 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.
Capstone D - Browser Ops Dashboard¶
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: 1. user story definition, 2. baseline UI, 3. SQL-backed filters, 4. exports and freshness indicators, 5. 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.
Expected output¶
- A practical portfolio proving SME-level operational Python capability.
Break/fix drills¶
- 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.
Troubleshooting¶
- 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.
Mastery check¶
You are SME-ready when you can: - demo all capstones end-to-end, - explain architecture tradeoffs, - support and recover production-like failures.
Learning-style options (Play/Build/Dissect/Teach-back)¶
- 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.