← Course Catalogue Intermediate Level

⬆ Intermediate · Requires Foundation

Exposure Data Preparation
for Catastrophe Modelling

The bridge between theory and practice. Learn how to read, clean, geocode, and prepare a real portfolio for a vendor cat model run — and how to interpret, validate, and communicate what comes out. The practical skills every junior cat modeller needs on day one.

📋 Prerequisite: Catastrophe Modelling Fundamentals or equivalent experience

4Modules

14Lessons

~5hDuration

80%Pass Mark

Your Progress

0% complete 0 of 14 lessons Start Course →

What You Will Be Able to Do

✓Read and diagnose a Schedule of Values (SOV) from any cedant or client

✓Identify and prioritise COPE data gaps that most affect model results

✓Assess geocoding quality and understand what poor geocodes do to hazard scores

✓Prepare a clean input file for Touchstone (Verisk) or RiskLink (Moody's RMS)

✓Configure a model run with appropriate settings for your use case

✓Read and interpret OEP/AEP curves, AAL, and Event Loss Tables from model output

✓Conduct a structured QA audit of exposure data and model results

✓Communicate model results and their limitations clearly to non-technical audiences

📋

Module 01 — What Is Exposure Data?

The raw material — what exposure data is, where it comes from, and why its quality controls everything

3 lessons

Before you can run a cat model, you need to understand what you are putting into it. This module covers the foundational documents and data fields that form the exposure input — the Schedule of Values, COPE data, and Total Insured Value. Understanding what each field means and what goes wrong with it is the prerequisite for everything that follows.

1.1

The Schedule of Values What an SOV is, what it looks like, who creates it, and how to diagnose its quality at a glance

READING + QUIZ

○

1.2

COPE Data: Construction, Occupancy, Protection, Exposure The four fields that drive vulnerability curve selection — what each means and what missing data costs you

READING + QUIZ

○

1.3

Total Insured Value: The Most Important Number in the File Why TIV is the single largest source of systematic error — underinsurance, inflation, and how errors flow through to model output

READING + QUIZ

○

📍

Module 02 — Geocoding: Putting Your Portfolio on the Map

The most fundamental and most commonly missed technical step in exposure preparation

2 lessons

Every cat model assigns a hazard score based on where the risk is. Geocoding — converting a text address into geographic coordinates — determines how accurately the model places each location on the hazard surface. A poor geocode produces a wrong hazard score regardless of how good everything else in the file is. This module teaches you how geocoding works, what goes wrong, and how to fix it.

2.1

What Geocoding Is and Why It Determines Hazard The resolution hierarchy, what each geocoding level means for model accuracy, and how it differs by peril and by country

READING + QUIZ

○

2.2

Geocoding in Practice: Common Problems and How to Fix Them PO Boxes, incomplete addresses, mailing vs. risk location, batch tools, quality scoring, and international challenges

READING + QUIZ

○

⚙️

Module 03 — Running a Clean Portfolio Through a Vendor Model

The hands-on workflow from raw SOV to interpreted results — using Touchstone and RiskLink

4 lessons

This is the operational core of the course. Starting with an orientation to the two dominant vendor platforms, these lessons walk through the complete workflow: preparing the import file, configuring the model run, and interpreting the outputs. All four lessons reference a consistent fictional portfolio — the Meridian Commercial Property account — so you can follow the data from raw SOV to final EP curve.

3.1

The Two Major Vendor Platforms: Touchstone and RiskLink How Verisk and Moody's RMS platforms are structured, who uses them, and why results differ between vendors for the same portfolio

READING + QUIZ

○

3.2

Input File Preparation: Building a Valid Model Import File Field mapping, occupancy and construction code translation, pre-run QA checklist, and common import errors

READING + QUIZ

○

3.3

Model Settings and Run Configuration Loss types, coverage options, return period selection, demand surge, secondary uncertainty — what each setting does and when to use it

READING + QUIZ

○

3.4

Reading Model Output: EP Curves, AAL, and Event Loss Tables How to extract the numbers that matter for underwriting, reinsurance design, and regulatory reporting from a real model output file

READING + QUIZ

○

🎯

Module 04 — Quality, Validation & Communication

How to assess what your output is worth, flag its limits, and make it useful to decision-makers

4 lessons

Producing a model output is the starting point, not the end point. This module covers the professional wrapper that makes model results reliable and defensible: understanding sources of uncertainty, conducting a structured QA audit, communicating results to non-technical audiences, and positioning yourself for the evolving data landscape.

4.1

Model Uncertainty and the Limits of Your Output Hazard, vulnerability, financial, and data uncertainty — the four layers, how they compound, and how to present results with appropriate confidence

READING + QUIZ

○

4.2

Exposure Data Auditing: A Practical QA Framework A reusable pre-run and post-run audit process: completeness scoring, red flag patterns, the data quality matrix, and how to document your findings

READING + QUIZ

○

4.3

Communicating Results to Non-Technical Audiences What underwriters, and clients need from a model run — and how to give it to them without losing the nuance

READING + QUIZ

○