MC-4824

Add Formula Data Analysis-style charts to the F1 predictor dashboard and validate how much FP analysis helps predictions. Requested by Elmar from Telegram. Scope: 1. Dashboard charts: - Add visual charts for FP analysis where data exists: weighted FP pace, ideal lap, long-run pace, tyre degradation, corner speed, load factor/lateral grip, aero/PU triangulation. - Dashboard should handle FastF1 deep data vs OpenF1 fallback gracefully: show richer telemetry charts only when rows exist; show explanatory fallback text otherwise. - Keep tables if useful, but add chart visuals rather than table-only views. 2. Confirm model integration: - Verify/document that FP/F1DataAnalysis-style signals are consumed in prediction scoring, especially pipeline/run.py -> signals/fp_pace.json/aero_pu.json -> pipeline/05_model.py. - Add/adjust tests if needed so these signals cannot silently stop affecting scoring. 3. Actual race telemetry validation: - For Monaco first, compare free-practice analysis signals against actual race telemetry/results. - Measure whether FP signals were directionally on target: FP pace vs race pace, ideal/corner/load/aero signals vs actual race performance where telemetry permits. - Summarize misses and recommended weighting adjustments. 4. Season backfill plan with FastF1 rate-limit safety: - Do not query the whole season in one run. - Add a rate-limited/backoff-capable script or documented workflow to backfill one race at a time, preferably using cached existing data first. - Propose which completed races to run next after Monaco. Acceptance criteria: - Tests pass. - Dashboard verified locally or via available route. - At least Monaco validation report produced. - Clear answer to Elmar on whether current FP analysis changed predictions and whether weights need adjustment. Expected check-in: today after first implementation/validation pass.