Meeting Minutes

Sprint 1 review

Architecture diagram walkthrough

Stack decisions presentation

Data structure analysis

Presentation of Star Schema ( sort of star )

Working agreements — DoD, DoR & Charter

Sprint 2 planning

[Dehlya] Sprint 2 demo — Bronze + Silver pipeline (15M+ rows ingested)

[Dehlya] Watcher optimization demo — pipeline cycles from 5min to 6s

[Dehlya] Documentation review — SETUP.md, ETL.md, ARCHITECTURE.md

[Dehlya] Architecture diagram updated

[Dehlya] Star schema v2 — changes based on teacher feedback

[Johann] Dashboard mockups presentation

[Sacha] Weather CSV status (#7, #8)

Sprint 3 planning — Gold layer & ML

[Dehlya] create_silver.py — auto DB creation + admin privileges (DB_ADMIN_URL)

[Dehlya] Watcher fix — date-based comparison (DD.MM sorting bug) + --scan flag

[Dehlya] Dashboard mockups — tenant view + admin view

[Dehlya] Code review — Sacha's weather_download.py PR

[Johann] ML — presence prediction model using motion sensors (model choice workflow on KNIME)

[Johann] Dashboard mockups

[Johann] SCRUM — sprint management

[Sacha] weather_download.py — PR submitted, feedback applied, pending final fixes

BLOCKER: Gold layer blocked — star schema v2 sent to Cosette, no 2nd feedback yet

Discussion: site documentation update (per-flow docs, ML architecture)

Sprint 4 planning

[Sacha] clean_weather.py — implementation (code review pending)

[Dehlya] Gold layer — implementation done

[Dehlya] Energy price decision — Oiken tariffs selected (both apartments on Oiken network)

[Dehlya] Code review for Sacha's Bronze-to-Silver work

[Johann] ML presence model — optimisation

[Johann] Research on writing predictions back to DB

[Johann] Scrum management updates

Sprint 4 planning

Energy pricing uses Oiken tariffs — both apartments are on Oiken's network

[Dehlya] Code reviews — Sacha's code

[Dehlya] Gold layer progress — fact tables implementation

[Sacha] clean_weather.py — updated after code review feedback

[Sacha] sFTP folder selection issue — date/path problems

[Johann] ML progress — presence & energy prediction models

Discussion: prediction duplicates strategy — keep history vs overwrite

Discussion: date formatting issues in weather pipeline

Sprint 5 planning

Prediction duplicates — keep history for now (no overwrite on re-run)

[Dehlya] Gold layer landed — OLAP model (#15), Gold DB (#16), Silver→Gold flow (#17)

[Dehlya] create_gold.py / populate_dimensions.py / populate_sensors.py / populate_weather.py merged

[Sacha] Weather Sources→Bronze (#7a) and Bronze→Silver (#7b) — both shipped

[Sacha] Slow-flow scheduler retrieving daily weather (#11b)

[Johann] KNIME — workflow scaffolding for presence prediction (#26a) and energy prediction (#27a + #27b)

[Johann] Mockups → Power BI dashboard prep

[Dehlya] Documentation refresh — per-flow docs on the website

Sprint planning — KNIME predictions back to DB, Power BI dashboards, RLS

Bronze→Silver weather pipeline approved as-is, ready for production

KNIME workflows will use Variable→Credentials pattern for runtime credential injection

[Johann] Power BI energy consumption dashboard (#19) — done

[Johann] Power BI environment dashboard — temp/humidity/CO₂/door+window (#20) — done

[Johann] Predictions written back to DB (#28) — gold.fact_prediction_motion / fact_prediction_consumption

[Johann] Prediction visualisation dashboard (#29)

[Dehlya] KNIME data export/import flow (#25) — both directions wired

[Dehlya] Presence model (#26) and consumption model (#27) running headless via run_knime_predictions.py

[Dehlya] Power BI row-level security per apartment (#24)

Discussion: scalability forecast structure (for #31)

Discussion: GDPR/ethics writeup (for #32)

Sprint planning — anonymisation, GDPR deliverable, scalability writeup, packaging

RLS implemented per-apartment using dim_apartment.apartment_id

Predictions kept as history (no overwrite) — confirmed from earlier sprint decision

[Dehlya] Data masking / anonymisation (#18) — building_name → 'Building <id>', owner_user_id stripped at silver→gold boundary

[Dehlya] GDPR & ethics written assessment (#32) — submitted

[Dehlya] Scalability forecast (#31) — projected storage & compute growth submitted

[Dehlya] Full data-flow scheduling (#11) — fast flow, slow flow, gold refresh, daily batch all wired into one watcher process

[Dehlya] Self-contained customer deployment package (#33) — installer tested end-to-end on the VM

[Dehlya] Compress-after-silver — 10-15× bronze shrink while preserving audit trail

[Dehlya] Postgres tuning + drop-constraint backfill — first-install silver from 4 h to ~6 min on re-run

[Sacha] Power BI filters added across the dashboards

[Sacha] User guide written

[Johann] Scrum management and project management throughout the sprint

[Johann] Documentation contributions and review

[Johann] Presentation slides — structure and content support

Discussion: defense slides — split per author; Bronze→Gold deep-dive owned by Dehlya (slides 7–13)

Discussion: technical doc, installation guide, user guide — all polished and exported to .docx

Final week — defense prep, presentation rehearsal, last documentation polish

First names retained as RLS pseudonyms at 2-tenant scale (documented in ADR-005)

SAP SAC track (#21, #22) abandoned — focus on Power BI for the defense

Storage encryption (#9), monitoring/alerts (#10), full IAM (#23), i18n (#30) deferred — out of scope for the school deployment, documented as future work