Hugging Face's logo

Datasets:
optimalcover
/
vsc-pricing

Tasks:
Tabular Regression
Tabular Classification
Modalities:
Tabular
Text
Formats:
parquet
Languages:
English
Size:
< 1K
Tags:
automotive
pricing
consumer-finance
insurance
warranty
tabular
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
vsc-pricing / README.md
anas-muhammad's picture
anas-muhammad
Publish schema v3.6.7
fbc9eec verified
preview code
|
raw
history blame contribute delete
5.83 kB
metadata 
license: cc-by-4.0
pretty_name: OptimalCover VSC Pricing Dataset
language:
  - en
size_categories:
  - n<1K
task_categories:
  - tabular-regression
  - tabular-classification
tags:
  - automotive
  - pricing
  - consumer-finance
  - insurance
  - warranty
  - tabular
  - benchmark
configs:
  - config_name: rates
    data_files:
      - split: train
        path: rates.parquet
  - config_name: eligibility
    data_files:
      - split: train
        path: eligibility.parquet
  - config_name: mileage_bands
    data_files:
      - split: train
        path: mileage_bands.parquet
  - config_name: pricing_config
    data_files:
      - split: train
        path: pricing_config.parquet

OptimalCover VSC Pricing Dataset

Independent, open reference data for Vehicle Service Contract (VSC / extended warranty) pricing — derived from public insurance filings and actuarial reserve disclosures.

Dataset summary

OptimalCover publishes pricing benchmarks for Vehicle Service Contracts based on public insurance filings, actuarial reserve disclosures, and documented industry cost structures. This is the Hugging Face mirror of the canonical GitHub release.

Configurations

Config Rows Description
rates 180 Canonical pricing rates by vehicle class × mileage band × term × deductible
eligibility 104 Make/model → vehicle class (A-D) mapping with eligibility flags
mileage_bands 4 Mileage band definitions used by the pricing model
pricing_config 5 Deductible factors, schema version, methodology constants

Usage 

from datasets import load_dataset

rates = load_dataset("optimalcover/vsc-pricing", "rates", split="train")
print(rates[0])
# {'vehicle_class': 'A', 'mileage_band': '...', 'term_months': 36, ...}

eligibility = load_dataset("optimalcover/vsc-pricing", "eligibility", split="train")
mileage_bands = load_dataset("optimalcover/vsc-pricing", "mileage_bands", split="train")
pricing_config = load_dataset("optimalcover/vsc-pricing", "pricing_config", split="train")

Data fields

rates

  • coverage_level (string) — always EXCLUSIONARY in v3.6.7
  • vehicle_class (string) — A, B, C, or D
  • mileage_band (string) — human-readable band label
  • term (string) — e.g. 48/60000 (months/miles)
  • term_months (int)
  • term_miles (int)
  • deductible (int) — always 100 in v3.6.7 (canonical)
  • band (string) — REFERENCE_LOW or REFERENCE_HIGH
  • price (int) — USD

Pair REFERENCE_LOW and REFERENCE_HIGH rows on the same (vehicle_class, mileage_band, term, deductible) to form a price range.

eligibility

Make/model mapping to risk class A-D.

mileage_bands

Raw band definitions — odometer min/max and inclusivity.

pricing_config

Encoded methodology constants. config_value is a JSON string; parse it to access nested rules (e.g. deductible multipliers, Class D cap).

Derived deductibles

Only the canonical $100 deductible is stored. $0 and $250 are derived at runtime from pricing_config.deductible_handling:

  • $0 = $100 × 1.35, rounded to nearest $50, capped at +$500 for Class D
  • $250 = $100 × 0.85, rounded to nearest $50

A worked notebook is available in the upstream GitHub repo at examples/03_deductible_comparison.ipynb.

Methodology

Full methodology lives at https://optimalcover.com/methodology. A condensed version ships alongside the data as methodology_summary.md.

Inputs include:

  • Public insurance filings (state DOI rate filings, NAIC disclosures)
  • Actuarial reserve disclosures from VSC underwriters
  • Documented dealer / admin / reserve cost structures
  • Vehicle reliability and parts-cost data

Outputs are reference price ranges (REFERENCE_LOW / REFERENCE_HIGH) — never single point prices, never quotes.

What this dataset is NOT

  • Not a quote, offer, or sale of any product
  • Not affiliated with any single VSC provider
  • Does not contain customer data, VINs, or any personally identifiable information

Schema versioning

Releases are tagged by the OptimalCover pricing schema version (e.g. v3.6.7). Each schema version is an immutable snapshot — once published, its data files do not change. Methodology refinements ship as new schema versions with a CHANGELOG.md entry.

Citation

Please cite this dataset as:

OptimalCover (2026). VSC Pricing Dataset (v3.6.7).
https://github.com/Optimal-Cover/vsc-pricing-data
Licensed under CC BY 4.0.

BibTeX:

@dataset{optimalcover_vsc_2026,
  author    = {{OptimalCover}},
  title     = {VSC Pricing Dataset},
  version   = {3.6.7},
  year      = {2026},
  url       = {https://github.com/Optimal-Cover/vsc-pricing-data},
  license   = {CC-BY-4.0}
}

License

This dataset is released under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. You are free to share and adapt the data for any purpose, including commercially, as long as you provide appropriate attribution.

Links