actuarialpy

actuarialpy

Purpose-neutral actuarial calculation primitives, plus the shared actuarial data contract — the foundation of the OpenActuarial ecosystem.

actuarialpy.Experience is the ecosystem's canonical semantic wrapper for historical actuarial data: it binds column roles, grain metadata, and snapshot context. Its domain operations are immutable transformations; calculations and workflow outputs belong to consuming packages.

Overview

actuarialpy provides the atomic building blocks the rest of the ecosystem is written against: ratios and per-exposure metrics, claim development and completion, trend fitting and projection, credibility, large-claim pooling, and financial mathematics. Everything operates on plain floats, NumPy arrays, and pandas objects, with a consistent type-mirroring convention (scalar in, float out; Series in, Series out with the index preserved).

The package deliberately contains no workflow orchestration and no domain-specific vocabulary — those belong to the workflow packages built on top of it. If a function here needs to know why you are calling it, it does not belong here.

Installation

pip install actuarialpy

Requires Python 3.10 or newer.

Quick start

import pandas as pd
import actuarialpy as ap

# ratios and per-exposure rates on any aggregate
print(ap.loss_ratio(1_240_000, 1_500_000))       # 0.8267
print(ap.per_exposure(1_240_000, 12_000))        # 103.33 per exposure unit

# trend claim severity to project future loss costs
monthly = pd.DataFrame({
    "month": pd.date_range("2024-01-01", periods=24, freq="MS"),
    "avg_severity": [5_000 * 1.004 ** i for i in range(24)],
    "claim_count": [20] * 24,
})
severity_trend = ap.fit_trend(monthly, date_col="month", value_col="avg_severity")
# if severity trends at +0.4%/month and claim count stays flat,
# projected losses next quarter will be:
projected_severity = ap.project_forward(monthly["avg_severity"].iloc[-1], 
                                        severity_trend.annual_trend, months=3)
projected_losses = projected_severity * monthly["claim_count"].iloc[-1]

# cap large claims at a pooling point; the excess moves to its own column
claims = pd.DataFrame({"member": ["a", "b", "c"],
                       "paid": [612_000.0, 340_000.0, 96_500.0]})
pooled = ap.pool_losses(claims, loss_col="paid", pooling_point=250_000)
print(pooled)

What's inside

• Metrics — loss/expense ratios, per-exposure rates, weighted statistics, contribution and comparison helpers.
• Reserving — completion triangles, chain-ladder development factors, Mack standard errors, completion applied back to tidy data.
• Trend and seasonality — trend fitting, forward projection, seasonal adjustment.
• Credibility — Bühlmann, Bühlmann–Straub, and limited-fluctuation credibility.
• Pooling — large-claim capping and excess extraction.
• Financial — time-value-of-money primitives (present/future value, annuities, rate conversions).
• Data utilities — exposure handling, banding, period alignment, member lifecycle status, margins and adjustments.

The full API reference and end-to-end worked examples live at openactuarial.org/actuarialpy.html.

The OpenActuarial ecosystem

actuarialpy is one of seven packages that share conventions — tidy tables, explicit distribution parameterizations, reproducible random-number handling — and compose across package seams:

Install everything at once with pip install openactuarial.

Development

git clone https://github.com/OpenActuarial/actuarialpy
cd actuarialpy
python -m pip install -e ".[dev]"
pytest
ruff check src tests

CI runs the same gate on Python 3.10–3.14 across Linux and Windows.

Versioning and stability

All ecosystem packages are pre-1.0: minor releases may change APIs, and every release is documented in CHANGELOG.md. Current per-package API stability is tracked at openactuarial.org/stability.html.

License

MIT — see LICENSE.