Relativities

Relativities is Effective’s Python package for insurance pricing. Build and test your own raters, explore what-if scenarios, and measure rate changes by segment. Run your programs and market programs against the same risks to investigate pricing differences. Rating runs on your machine. Your submission data and results stay local. Install Relativities and set your API key to get started.

from relativities import Program

program = Program.load("effective://example-ca-ppa")
book = program.example_book
results = program.rate(book)

results[["submission_id", "status", "premium"]].head(3)
#   submission_id     status premium
# 0       S000001  succeeded    2636
# 1       S000002  succeeded    5339
# 2       S000003  succeeded    2341

This synthetic California personal auto program includes an example book of 100,000 policies. Its premium outcome is the annual premium. Each submission keeps its result row, status, and reason. Your policy data and rating results stay local.

Core tasks

Task	What you get
Create a book	Your policies, checked against the line’s schema
Download a program	A pinned program in a local folder
Load a program	A program with its schema, inputs, and tables
Rate a book with a program	One row per submission, with its status and reason
Create a market	Named programs to compare
Rate a book against a market	Results for each submission and program
Create a study	Scenarios and sampled versions of your book
Rate a study with a program	Results for each submission, scenario, and draw
Rate a study against a market	Study results across programs

How it works

A book holds your policy facts in pandas tables. A program prepares those facts for its rating algorithm and calculates the results. A market compares several programs. A study creates scenarios and sampled draws from a book to test changes and assumptions. Rating a book or study returns a DataFrame. Use pandas to analyze the results, and explain to inspect the calculations behind a premium.

Build your own rating program

Use RSL to write rating calculations, input constraints, tables, rounding, and source citations. Package a program to bundle the rater with its input preparation function, schema, and examples.

Get started

Rate policies

Compare programs

Explore scenarios

Build and test raters

Concepts

Python reference

Data reference

Core tasks

How it works

Build your own rating program

​Core tasks

​How it works

​Build your own rating program

Core tasks

How it works

Build your own rating program