Management Actions

Push a Loss Event (in-browser, demonstrates the ingest flow)

In the real system this is POST /api/v1/loss-events. Here it just appends to the in-memory dataset — click Recalculate afterward to see the dashboard respond.

Batch Upload (CSV or Excel)

In the real system this is POST /api/v1/loss-events/upload; here it's parsed entirely in your browser. Accepts .csv, .xlsx, .xls. Headers can be snake_case (loss_date, gross_loss_amount, ...) or the companion Excel model's column names (Loss_Date, Gross_Loss, ...) — case-insensitive either way. Download a blank template. CSV parsing has no dependencies. Excel parsing loads a small parser library (SheetJS) from a CDN the first time you pick an .xlsx file — the only network request this page ever makes, and only if you use that feature; CSV works fully offline.

Configuration

Edit and click Recalculate — every threshold below drives the RAG colors and actions above, live.

How this demo differs from the real API

This file:            JS Monte Carlo engine + embedded 300-event sample, runs on page load.
Full SaaS system:     Flask REST API + SQLite + real push/pull endpoints + webhooks + auth.

Same math in both:    Poisson(lambda) frequency, Lognormal(mu,sigma) severity, per business line,
                      fit by method of moments; exact per-event compound simulation (no
                      Fenton-Wilkinson approximation) — this demo runs the same algorithm as
                      engine.py in the SaaS package, just ported to JavaScript.

To see the real, callable version with authentication, webhooks, and a persistent database,
run the SaaS package: python3 seed.py && python3 app.py  (see its README.md).