Bermuda

Bermuda is a Python package for the representation, manipulation, and exploration of insurance loss triangles, created by Ledger Investing. It offers a user-friendly interface for:

• Loading and saving insurance loss triangles using a number of formats (e.g. JSON, CSV, Pandas DataFrame objects, binary files).

• A single Triangle class for manipulating triangles of varying complexities (e.g. ragged, multi-program, cumulative or incremental triangles).

• An intuitive Cell type that can hold multiple data types and metadata.

• A collection of useful Cell- and Triangle-level functionality, including summarizing, aggregating, extending, filtering, and bootstrapping.

• A variety of plots to explore triangle structures, loss development dynamics, and predictive distributions and uncertainty, using Altair.

If you’re new to Bermuda, take a look at the Quick Start guide for a brief overview of its functionality, or the User Guide for a more complete explanation of Bermuda’s design decisions, insurance triangles, and Bermuda’s overall architecture. The Tutorial section includes common usage patterns.

If you’re interested in contributing to Bermuda, take a look at our Developer Guide.

• Installation
• Quick Start
• User Guide
  • Bermuda’s Scope
  • Data Structure Design
  • Triangles
  • Architecture
• Tutorials
  • Common triangle manipulations
  • Triangle Input/Output
  • Deriving key triangle quantities
  • Plotting triangle data
• Developer Guide
  • Contributing
• API
  • Triangle class
  • io functions
  • Cell and Metadata classes
  • utils functions
  • date_utils
  • matrix data types

Indices

• Index

• Module Index

• Search Page