v2.7.0 — Teaching Layers epic

Structure that turns eighty demos into a course.

• #114 Capstone — one scored scenario across the whole lifecycle: five leadership decisions, graded with reasons
• #115 Role tracks — curated playlists for Founder/CEO, Clinician/Researcher, Product Manager, Consultant/Analyst (links validated by test)
• #116 Quizzes per epic — quick knowledge checks with instant feedback across six epics
• #117 Bilingual glossary — 35 terms in English + العربية, searchable in either language
• #118 Case studies — Target, Zillow, Amazon, COMPAS, Netflix, Cambridge Analytica as sourced cards

This completes the second feature wave (Judgment, GenAI, XAI, Advanced Analytics, Teaching Layers) — 68/68 modules across 16 epics. Offline suite: 338 passed, 97% coverage.

v2.6.0 — Advanced Analytics epic

Optimization, simulation, geography, networks, and deeper recommendations — the analytics that turn data into decisions.

• #108 Price optimization — find the profit peak, not the highest price or the most units
• #109 Monte Carlo — run the future 10,000 times for a range and a probability of loss, not one fragile number
• #110 Geospatial — Nour Store orders on the map of Syria, sized by volume
• #111 Network analysis — referral influencers + community detection via dependency-free label propagation
• #112 Matrix-factorization recommender — NMF learns hidden tastes and fills the gaps in the purchase grid

All dependency-light (scikit-learn, numpy, graphviz, st.map) and deterministic. Offline suite: 311 passed, 97% coverage.

v2.5.0 — Explainability (XAI) epic

Open the black box: two honest, model-agnostic techniques on the real Predict model, no heavy SHAP dependency.

• #105 Feature attribution — global permutation importance: shuffle each input, measure the accuracy drop, and see what the model truly relies on (and why permutation beats impurity importance).
• #106 Single-prediction explanation — occlusion from a typical-customer baseline turns one prediction into plain-language reasons and a per-feature push chart — a defensible per-decision explanation.

Both are labeled as approximations that ignore some feature interactions. Offline suite stays green at ~96% coverage.

v2.4.0 — GenAI / LLM epic

The GenAI / LLM layer everyone asks about — six demos, each with a free-provider live path (Groq / OpenRouter / Hugging Face / local Ollama, via their OpenAI-compatible endpoints) and an honest offline simulation so the whole epic runs on stage with no network and no paid key.

• #98 Prompt playground — one task, four prompts (vague → engineered), scored on a 5-point rubric
• #99 RAG — retrieve from a company knowledge base, answer grounded, refuse when out of scope
• #100 Summarization — extractive offline / abstractive live, with the trade-off spelled out
• #101 Structured extraction — messy messages → a clean table with rules, model for the messy rest
• #102 Zero-shot classifier — route tickets into named categories with no training data
• #103 Guardrails & hallucination — why models invent facts, and a grounding fact-checker

Shared workshop/llm.py auto-detects the provider and labels every answer's source in the UI, so a simulation is never passed off as a real model. Full offline suite: 277 passed, 96% coverage.

v2.3.0 — Data Judgment epic

The leader-judgment core — how data lies:

• Simpson's paradox
• Correlation vs causation (confounders)
• Sampling bias
• Misleading vs honest charts
• Confidence intervals & p-hacking

233 tests passing; 67 pages.

v2.2.0 — Platform & UX epic

Turns the demo set into a polished product:

• Bilingual English/Arabic toggle (RTL)
• Bring Your Own Data (CSV upload → profile + chart)
• Presenter mode (agenda, talk-track cues, countdown timer)
• Journey / progress tracker
• Neurobotics theming (brand colours, logo, footer)
• Export my session (zip of report + cleaned data)

228 tests passing; 62 pages.

v2.1.0 — Engineering & Delivery epic

Repo maturity + delivery:

• CI on every PR (required status check) with enforced test coverage (≥90%, currently 98%)
• Dockerfile, Streamlit config, DEPLOY.md, and an in-app QR share page
• CONTRIBUTING guide and issue/PR templates

218 tests passing; 57 pages.

v2.0.1 — CI & reproducibility

• GitHub Actions CI runs the offline test suite on every push and PR (now a required status check on main)
• Added missing lxml dependency (caught by CI) so the suite passes on a clean environment

v2.0.0 — Complete Data Lab (35 modules, 9 epics)

The full Data Lab roadmap is complete: 35 interactive modules across 9 lifecycle stages, each developed on its own branch and merged via pull request.

Acquire · Label · Clean · Encode · Mine · Model · Scale · Govern · Serve

• 54 pages (6 slides + 48 demos), one shared 'Nour Store' story
• 216 tests passing; pure logic unit-tested, every page smoke-tested
• Colourblind-safe visuals, honest framing throughout (no faked GANs/Spark/trained models)

Built for the From Data to Decisions workshop by Neurobotics Academy.

v1.5.0 — Govern / Trust epic

Adds the Govern / Trust epic:

• PII masking / anonymization
• Bias / fairness audit
• Data drift monitoring (PSI)
• Data lineage / catalog / versioning

209 tests passing; 47 pages. One epic left in Issues: Serve.