Skip to content

Public demo parity with the secure dashboard

Ticket #293: Public demo parity with the secure dashboard
Type: Feature / User Experience / Business Analysis
Affected Component: code_source_simule/flask_app.py, templates/dashboard.html, templates/demo_index.html, templates/titre_detail.html, tests/test_dashboard.py, tests/test_demo.py, docs/tests/, docs/fr/tests/, specs/ajout_indicateurs_demo/

1. Context

This report directly follows my last three indicator deliveries: management rules indicators, analyst signals indicators, and P/E valuation indicators. The vision remains unchanged: turn the dashboard into a real decision-support tool and prepare a clear foundation for future advisory AI (see Project Vision).

Until now, the public showcase of the app (the demo, accessible without login) was significantly weaker than the secure version. It did not reflect the true product depth, which reduced the impact of the demo for anyone discovering the app. At the same time, the secure dashboard had accumulated some blocks that had become redundant over time.

2. Objective

Align the public demo with the secure dashboard so it becomes an almost perfect reflection, while simplifying the secure dashboard by removing blocks that no longer provide value.

Concretely, the goal was to:

  • remove only the obsolete blocks from the secure dashboard, without touching previously validated business cards;
  • reproduce in the demo the same card selection and the same visual language as the secure dashboard;
  • keep the demo complete and credible even when real data is insufficient, without ever exposing sensitive information.

3. What Was Delivered

  • A cleaner secure dashboard: removal of redundant blocks (recommended decisions, weighted P/E, and legacy summary counters), while preserving all useful business cards (gains to secure, titles to reinforce or liquidate, analyst targets, P/E valuation, contrary-to-expectations signals, currency exposure, portfolio performance, and 52-week analysis).
  • A demo aligned with secure behavior: the public page now uses the same cards, the same titles, and the same presentation, resulting in a nearly identical experience to the authenticated version, minus the login gate.
  • A complete but safe demo: cards remain populated with a coherent, stable demo dataset, under strict confidentiality constraints: up to ten titles and one coherent snapshot only, without exposing sensitive real data.
  • An enriched demo holding-detail page: the "Analyst Targets" gauge is now available on the demo detail page too, matching the secure experience.
  • Stronger test coverage: six new demo scenarios were added (visual parity, shared card reuse, secure/demo coherence, confidentiality cap, single snapshot, and reproducibility), and FR/EN test documentation was aligned.

4. Business Impact

  • A credible product showcase: the public demo now reflects the true maturity of the product, improving first impressions.
  • Sharper readability on the secure side: removing non-essential blocks reduces visual load and keeps focus on the indicators that matter most.
  • Controlled confidentiality: the demo remains complete and convincing without exposing personal or sensitive data.
  • Vision continuity: this parity creates a more homogeneous foundation for future advisory AI, based on the same card logic on both sides.

5. Validation and Status

  • Full test suite green: 211 passed.
  • Key behaviors covered by dedicated tests: removed-block disappearance on secure dashboard, preservation of validated business cards, card-title parity between demo and secure views, demo confidentiality cap (maximum ten titles with bounded history), single coherent snapshot, and deterministic demo dataset reproducibility.
  • Test documentation updated: demo report now at 83% coverage, dashboard report at 91%, and overall documented coverage recalculated to 80%.
  • One inherited obsolete test case was renamed and realigned with the simplified dashboard behavior to avoid future ambiguity.
  • Traceability: work delivered on branch ajout_indicateurs_demo and scoped by specs/ajout_indicateurs_demo/.

6. Lessons Learned

  1. A public showcase only has value when it is faithful: aligning demo and real behavior is far better than maintaining a simplified demo that undersells the product.
  2. Simplification is a quality decision on its own: removing a redundant block can improve the experience as much as adding a new one.
  3. Token usage summary: For this implementation, the recorded token consumption was $8.55 for 855.49 Copilot AI credits. Compared to the three recent implementations mentioned above, this release cost approximately 10.75% less than adding P/E indicators (3rd implementation), 23.52% less than analyst signals (2nd implementation), and 48.09% less than business rules (1st implementation); this confirms and amplifies the optimization trend already underway.