Wire demand_weather_forecast.json into augur (consume + visualise)

[ducroq]

Context

energydatahub publishes `demand_weather_forecast.json` daily (~462 KB) from `OpenMeteoWeatherCollector` at 11 NL/DE/BE population centers (Amsterdam, Rotterdam, Den Haag, Utrecht, Eindhoven, Groningen, Düsseldorf, Köln, Hamburg, Brussels, Antwerp). It includes temperature, apparent_temperature, humidity, precipitation, wind, cloud cover, and pre-computed Heating/Cooling Degree Days (HDD/CDD).

Today augur does NOT consume this file — it is not in `DATA_FILES` in `decrypt_data_cached.py` and the live URL 404s. The data is being produced and going to waste.

Why this matters

Two distinct opportunities, each could justify its own follow-on:

1. Dashboard: a Demand tab showing population-weighted HDD/CDD or temperature trends across CWE — useful market context for users reading wholesale price charts. The data is already there.
2. Model features (separate experiment, augur#19 umbrella): HDD/CDD are well-established demand-driving features for short-horizon electricity price forecasting. Not currently in the LightGBM-Quantile feature set (`ml/shadow/features_pandas.py`). Worth an EXP under augur#19 calibration follow-on, once calibration is sorted.

Scope of this issue

Only (1) — the dashboard wire-up. Adding HDD/CDD as a model feature is a separate experiment per ADR-007 (pre-committed criterion, A/B against current 24-feature set).

Concrete plan

1. Add `'demand_weather_forecast.json'` to `DATA_FILES` in `decrypt_data_cached.py` so Netlify decrypts it
2. Extend `static/js/modules/tab-charts.js` with a new `processDemandWeather` function reading `temperature` / `hdd` / `cdd` per location
3. Add a small chart block under the existing Weather tab, or a new Demand tab — design call
4. Document the field schema in CLAUDE.md Key Paths

Dependencies

• Soft prerequisite: weather-pipeline-unification work (in progress 2026-06-05 in energydatahub repo). Doesn't block this, just want the upstream collector enriched first.

Out of scope

• Wiring HDD/CDD as a LightGBM feature — file under augur#19 EXP-018+ when calibration is resolved
• Historical / archive demand-weather data — separate concern (see augur#21)

🤖 Generated with Claude Code

[ducroq]

Cross-reference: the model-feature framing for the same demand_weather_forecast.json data source is now tracked separately at #22 (HDD/CDD as long-horizon demand features for h+25..h+72, EXP-018 candidate). This issue (#20) stays scoped to the dashboard wire-up. They are independent — one or both can land; neither supersedes the other.