inst/database/README.md

In jimbrig/lossrx: Actuarial Loss Development and Reserving with R

Database for lossrx Package: actuarialdb

TLDR

Get up-and-running by simply entering this folder and running docker-compose up --build to build and run a container instance of the actuarialdb locally on port 5432.

For reference here is the docker-compose.yml file:

services:
    postgres:
        image: postgres:latest
        restart: always
        environment:
            POSTGRES_USER: postgres
            POSTGRES_PASSWORD: admin
            POSTGRES_DB: postgres
        logging:
            options:
                max-size: 10m
                max-file: "3"
        ports:
            - '5432:5432'
        volumes:
            - ./postgres-data:/var/lib/postgresql/data
            - ./SQL:/docker-entrypoint-initdb.d

Structure

> tree inst/database

./inst/database
├── db.dbml
├── db.sql
├── Dockerfile
├── docker-compose-yml
├── package.json
├── config.yml
├── README.md
├── .vscode/
├── postgres-data/
├── backups/
├── scripts/
├── docs/
├── CSV/
├── SQL/
└── YAML/

8 directories, 7 files

Root Level Files

• db.sql: SQL script to generate database schema - autogenerated using DBML (Database Markup Language) Command Line Tool by running dbml2sql db.dbml against the db.dbml markup file.

• db.dbml: DBML (Database Markup Language) script for converting SQL into documentation. I originally generated this document using dbdiagram.

• Dockerfile and [docker-compose.yml]: Used to create database PostgreSQL container instance.

• config.yml: Used by the R package config to load database connection credentials and gain access to the database from R.

postgres-data

This directory is mapped to the docker containers volume as a locally mounted docker volume to store postgresql data related to the database and persist this data across container runtimes. Most of this directories contents are git ignored, however I left some configs for observational purposes.

Scripts

The scripts directory houses various utility .sh, .R, .ps1, etc. scripts or queries related to our database.

Backups

The backup directory houses any snapshots or database backup files.

SQL

SQL scripts to create database schema and tables.

CSV

CSV data files to seed data into the database tables.

YAML

YAML configuration files representing the database schema by table.

Design

Introduction

The actuarial database is meant to serve as an enterprise data model (EDM) which captures the semantics, operations, and relationships relevant to a Property Casualty Operating Business Model.

It is important to note the difference between Transactional vs. Analytical data warehouse implementations as well as how to deal with multi-dimensional online analytical processing (OLAP) data structures.

For example, a triangle which gets derived from a suite of lossruns is a structure representing claim values over time grouped by evaluation dates, accident/policy years, and coverage (with optional further segmentation i.e. limits, divisions, segments, lines of business, etc.). This is a multi-dimensional data series.

When dealing with multi-dimensional data structures the optimal data warehouse implementation is a dimension fact table design where a central claims fact table is supplemented by various dimension tables which contain claim attributes.

Inspiration

OMG's Property Casualty Data Model Specification

Among the components of the P&C data model, there are the entities, attributes and relationships.

The main objects of the data model are the Entities. An Entity represents a person, organization, place, thing, or concept of interest to the enterprise.

Next, an Entity can express a very broad and varied set of instance data, or it can be very specific. This is referred to as levels of abstraction.

Some examples of the core entities are:

− Insurable Object: An Insurable Object is an item which may be included or excluded from an insurance coverage or policy, either as the object for which possible damages of the object or loss of the object is insured, or as the object for which damages caused by the object are insured.

• Examples: residence, vehicle, class of employees. − Vehicle: A Vehicle is an Insured Object that is a conveyance for transporting people and/or goods. − Person: Person can be a human being, either alive or dead. − Organization: An Organization is a Party that is a business concern or a group of individuals that are systematically bound by a common purpose. Organizations may be legal entities in their own right.
• Examples: commercial organizations such as limited companies, publicly quoted multinationals, subsidiaries. − Agreement: Agreement is language that defines the terms and conditions of a legally binding contract among the identified parties, ordinarily leading to a contract.
• Examples; policy, reinsurance agreement, staff agreement. − Claim: Claim is a request for indemnification by an insurance company for loss incurred from an insured peril or hazard. − Claim Amount: Claim Amount is the money being paid or collected for settling a claim and paying the claimants, re-insurers, other insurers, and other interested parties. Claim amounts are classified by various attributes. − Policy Coverage Detail: Policy Coverage Detail defines the coverage's included in an insurance policy (refer to Coverage definition). It is a key link in the data model among Policy, Claim, and Reinsurance.

Relationships:

An Insurable Object is covered as defined in a Policy Detail by its Coverage (One to Many).

A Claim is settled and results in a Claim Amount

jimbrig/lossrx documentation built on Dec. 20, 2024, 3:46 a.m.