FAQ

An open-source classification system for federal contractors. It ingests five years of obligation data from USASpending.gov, applies a two-axis classification, and produces a Posture Class assignment (1-4), velocity label, proximity map, and public contractor dossier for every firm with base contract activity in a state.

Nevada is the initial coverage area with 779 classified contractors.

A four-class classification based on two axes: total contract dollars (base + delivery orders) and base contract count over a trailing 5-year window.

• Class 1 — High volume ($5M+) and high frequency (3+ contracts). Systematic winners.
• Class 2 — High volume ($5M+) but low frequency (<3 contracts). Concentrated risk.
• Class 3 — Low volume (<$5M) but high frequency (3+ contracts). Growth pipeline.
• Class 4 — Low volume (<$5M) and low frequency (<3 contracts). Entry level.

Thresholds are fixed ($5M volume, 3 contracts) and state-invariant. Only base contracts count: definitive contracts and purchase orders. BPA calls and delivery orders excluded.

Each contractor with 2+ base contracts gets a cadence-based velocity label measured against their own historical award rhythm:

• Accelerating — winning faster than their historical pace
• On pace — within their normal cadence
• Slowing — longer gaps between wins than before
• Declining — significant drop-off from historical rhythm
• Inactive — no awards in 5+ gap cycles

Single-contract contractors have no velocity to measure. Velocity appears on contractor dossiers and in the population-level carousel charts.

Total contract dollars divided by base contract count. Two contractors with identical volume but different Obligation Density have fundamentally different structures. One winning 200 contracts at $500K each vs one winning 2 contracts at $50M each. A single-axis revenue ranking conflates them. Obligation Density exposes the difference.

For each contractor, the six firms in the same state that win the most similar contracts at the most similar scale. Derived from awarded contracts, not SAM.gov registration claims.

Code overlap: each shared NAICS or PSC code weighted by 1/n² where n = contractors holding that code. PSC codes carry 2x base weight. Rare code overlaps dominate.

Scale filter: overlap multiplied by (min/max)² of base contract volumes. A 1:10 mismatch decays to 1%. A 1:100 decays to 0.01%.

The map is directional: Contractor A appearing in B's map does not guarantee the converse.

The public profile for a single firm at fedcompindex.org/dossier/[name]/. Displays: Posture Class, base contract volume, base contract count, obligation density, velocity, full contract history, proximity map, SBA certifications, and spectator count.

Each dossier generates a posture card image used as the OG/meta preview when the URL is shared. Dossier URLs are permanent.

SAM.gov is a registration ledger. USASpending.gov is a transaction ledger. Neither classifies contractors or maps procurement overlap.

FedComp Index joins both and applies a classification function and proximity algorithm that neither source provides. The underlying obligation data is identical to what any analyst can query from the public APIs. The classification layer (Posture Class, velocity, proximity map) is the differentiator.

The data archive of FedComp Index: datasets on HuggingFace and Kaggle (CSV, Parquet), classification engine on PyPI (fedcomp-index), pre-classified data bundles (fedcomp-index-data), and source repository on GitHub. All artifacts MIT-licensed.

• HuggingFace
• Kaggle
• PyPI: pip install fedcomp-index-data

Key columns: posture_class, total_dollars_5yr, base_contract_count, obligation_density, velocity_cadence, last_award_date, primary_naics. MIT licensed. Updated monthly.

Incremental obligation pulls from USASpending Monday through Saturday. Full re-ingestion Sundays. Classifications recalculate and the site regenerates after each pull. External mirrors (HuggingFace, Kaggle, PyPI) update monthly.

FAR Part 10 — Market Research. The proximity map provides a code-weighted, scale-filtered alternative to SAM.gov keyword search, surfacing firms by demonstrated obligation history under specific NAICS and PSC codes.

FAR 19.702 — Subcontracting Plans. Filtering the proximity map by certification type and Posture Class produces a set of firms with demonstrated contract performance in the relevant codes.

Teaming. The proximity map shows which Class 1 and Class 2 firms operate in the same NAICS codes as Class 3 and Class 4 firms at asymmetric scale. That scale asymmetry is the teaming aperture.