From portfolio data to decision intelligence.
Quantvesting is a quantitative portfolio/decision-support framework with an independent Python/Jupyter path and a Cloudflare-native Web assessment path. Both are intended to use the same portfolio, market-snapshot and methodology contracts.
ASSESS
↓
SEE THE SHAPE OF YOUR CAPITAL
↓
UNDERSTAND WHAT DESERVES ATTENTION
↓
KNOW WHERE TO RESEARCH NEXT
The customer-facing Web UI exposes Assess My Portfolio, About Quantvesting, How to Prepare & Upload, then My Portfolio, Review Areas, Opportunity Universe, and How to Read. My Journey remains an underlying Python/Web data capability but is intentionally not exposed in the first customer assessment.
The visual storytelling contract is Finding → Meaning → Evidence → Question → Context. The design intentionally prioritises clarity over dashboard density.
- Minimum-input portfolio onboarding:
Symbol, Shares, AvgCost - Broker/export aliases such as
Ticker,Quantity, andAverage Price - Portfolio health, allocation and concentration analysis
- Review areas and advisory capital-rotation review
- Quantvesting opportunity universe and ranking
- Longitudinal portfolio journey/EOD history
- Canonical market snapshots and Python ↔ Web parity fixtures
- Cloudflare-native background assessment processing
- Downloadable assessment PDFs generated on demand from the completed result
- Active customer notebooks plus historical notebook archive
Quantvesting is assessment/review oriented. It does not execute trades, manage client funds, guarantee returns, or expose internal action codes as customer-facing trading instructions.
The standard disclosure is maintained consistently across the active notebooks and Web UI. See the consolidated project document for the canonical wording and customer terminology.
quantvesting_v3/
├── README.md
├── docs/QUANTVESTING.md # consolidated active project context + roadmap
├── config/
├── market_data/
├── portfolio_data/
├── portfolio_template/
├── src/quantvesting/
├── notebooks/ # active customer + admin notebooks
├── scripts/
├── tests/
└── web/ # Cloudflare Worker + Web UI
Historical phase/changelog Markdown files have been consolidated into docs/QUANTVESTING.md. Historical notebooks remain under notebooks/archive/ because they are useful reference material.
After a Web assessment completes, the user can download a PDF copy of the assessment. The PDF is generated on demand from the completed assessment result and is not stored as a public report artifact.
The guest flow therefore does not expose a persistent bearer-style report URL. Assessment/job data needed for processing and the active registered-user workflow remains stored according to the existing repository model.
The print/PDF report includes My Portfolio, Review Areas and Opportunity Universe content plus the disclosure. How to Read, controls/guidance and My Journey are excluded from the customer-facing print report; the end-of-report assessment CTA, community CTA and disclosure are included.
WhatsApp is not implemented in the current release. It remains deliberately deferred until a suitable WhatsApp provider/API is selected. The future design keeps report persistence independent from WhatsApp and uses separate consent for report delivery and optional Weekend Quantvesting broadcasts.
See docs/QUANTVESTING.md → Roadmap → Deferred: WhatsApp.
From the repository root:
PYTHONPATH=src python -m pytest -q tests
npm run check:worker
npm run test:cloudflare
python scripts/check_web_cohesion.py
PYTHONPATH=src python -m pytest -q testsFor the full local regression/build sequence:
npm run test:allDo not claim a change is complete until the relevant regression tests pass.
The assessment is intentionally being treated as a financial story, not a dashboard. The user journey is:
See the shape of your capital
↓
Understand what deserves attention
↓
Know where to research next
Important presentation grammar:
Finding → Meaning → Evidence → Question → Context
The opening Start Here cards now surface the finding, the meaning/reason for attention, capital at stake and the question worth investigating. Review Areas and the Opportunity Universe continue the same narrative rather than introducing unrelated metrics.
Capital Rotation Review does not force a one-to-one mapping between a mature holding and a single alternative. When N rotation candidates are surfaced, the Web assessment presents up to N highest-ranked qualifying research opportunities as a reference opportunity set. Each rotation candidate is then compared conceptually with that set.
This avoids artificial precision such as Holding A → Opportunity A when the framework has no explicit matching rule. The section asks whether any of the reference opportunities may represent a better research case for the capital currently under review; it does not issue a sell/buy instruction.
The same reference-set concept is carried into the compatibility Worker PDF and the Python decision helper while preserving the existing first-reference fields for backward compatibility.
The product/assessment logic is deliberately being kept stable while the next operating phase focuses on distribution, real-user feedback, aha-moment discovery and willingness-to-pay evidence. Presentation improvements should make the existing analytical output easier to understand; they should not add dashboard density without evidence that it improves customer value.
Capital Rotation Review now uses unheld qualifying opportunities as its reference set. When N rotation candidates are surfaced, the Web/PDF presentation shows up to N highest-ranked qualifying opportunities that the user does not currently hold. This keeps the section focused on opportunity cost: the question is whether currently deployed capital deserves comparison with a stronger unheld research opportunity.
The existing adaptive N-sized reference set and no-forced-one-to-one mapping remain unchanged. If fewer than N qualifying unheld opportunities exist, the reference set is smaller rather than being padded with already-held securities.