Summary
Promote the MVP scope of docs/optional-features/03-scenario-simulation-what-if-planning.md into code via a new app/features/scenarios/ vertical slice. Turns ForecastLabAI from "predict the future" into "plan possible futures": run a baseline forecast from an existing trained model, apply deterministic, transparent uplift/drag factors for future assumptions (price change, promotion, holiday, inventory, lifecycle), and return a baseline-vs-scenario comparison.
Today every forecast answers only "what happens if nothing changes?" — there is no surface that answers "what if we discount this SKU 15% next week?"
Scope (phased — each phase independently shippable)
- Phase A — Stateless Simulation Engine (backend): deterministic
adjustments.py + stateless POST /scenarios/simulate — resolves a baseline model, runs a baseline forecast, applies per-day adjustment factors, returns a ScenarioComparison. No table.
- Phase B — Saved Scenario Plans (persistence): new
scenario_plan table + Alembic migration; POST /scenarios, GET /scenarios, GET /scenarios/{id}, DELETE /scenarios/{id} CRUD.
- Phase C — What-If Planner Page (frontend):
/visualize/planner — baseline picker → assumption form → run → baseline-vs-scenario chart + delta table → save/reload/delete named plans.
Out of scope
The "Full Version" (future-feature-frame generator, exogenous-regressor model support, agent-generated scenarios, multi-scenario comparison) — explicitly deferred.
Plan of record
- PRP:
PRPs/PRP-26-scenario-simulation-what-if-planning.md
- Feature brief:
docs/optional-features/03-scenario-simulation-what-if-planning.md
Builds on already-merged slices: forecasting (PRP-5), registry (PRP-7), jobs (PRP-8), data_platform (PRP-2), dashboard (PRP-11).
Summary
Promote the MVP scope of
docs/optional-features/03-scenario-simulation-what-if-planning.mdinto code via a newapp/features/scenarios/vertical slice. Turns ForecastLabAI from "predict the future" into "plan possible futures": run a baseline forecast from an existing trained model, apply deterministic, transparent uplift/drag factors for future assumptions (price change, promotion, holiday, inventory, lifecycle), and return a baseline-vs-scenario comparison.Today every forecast answers only "what happens if nothing changes?" — there is no surface that answers "what if we discount this SKU 15% next week?"
Scope (phased — each phase independently shippable)
adjustments.py+ statelessPOST /scenarios/simulate— resolves a baseline model, runs a baseline forecast, applies per-day adjustment factors, returns aScenarioComparison. No table.scenario_plantable + Alembic migration;POST /scenarios,GET /scenarios,GET /scenarios/{id},DELETE /scenarios/{id}CRUD./visualize/planner— baseline picker → assumption form → run → baseline-vs-scenario chart + delta table → save/reload/delete named plans.Out of scope
The "Full Version" (future-feature-frame generator, exogenous-regressor model support, agent-generated scenarios, multi-scenario comparison) — explicitly deferred.
Plan of record
PRPs/PRP-26-scenario-simulation-what-if-planning.mddocs/optional-features/03-scenario-simulation-what-if-planning.mdBuilds on already-merged slices:
forecasting(PRP-5),registry(PRP-7),jobs(PRP-8),data_platform(PRP-2), dashboard (PRP-11).