MachineryPete: Used Farm Equipment Marketplace Data

Machine-level records from MachineryPete, the used farm equipment marketplace operated inside the Farm Journal media portfolio. Every row is one unit of iron: roughly 16,000 used-tractor listings in the tractors category alone at research time, sitting inside a fourteen-category catalog that runs from harvesting, planting and applicators through hay and forage, tillage and grain handling to loaders, trucks and trailers, livestock equipment, specialty crops, construction and lawn and garden.

Each listing card carries a consistent eleven-field schema - year, make, model, asking price in dollars, engine hours, drive type, the seller's own condition note, city and state, seller name, photo count and category - and the detail layer beneath it adds serial numbers, stock numbers and model-specific specifications. Alongside the live inventory sits an Auction Price Data section and an Upcoming Auctions calendar, which is what makes this record the auction-grounded complement to pure asking-price feeds. Get a sample of this dataset cut to your makes, categories and states.

Two complete captures from the August 2026 research pass, exactly as they arrive - one row per unit, typed, with the seller's prose preserved rather than paraphrased:

# machinerypete listing rows - captured in the August 2026 research pass

title       : 2000 John Deere 7810          category : Tractors
price_usd   : 95900         hours : 4227    drive : MFWD
location    : Washington, MI                seller_type : Ask Owner
note        : Retirement Sale. One Owner, always stored inside.
              Front fenders, break away flashers, buddy seat.

title       : 2005 John Deere 8320          category : Tractors
price_usd   : 129900        hours : 3516    drive : MFWD
location    : Washington, MI                seller_type : Ask Owner
note        : Tractor is excellent. Always stored inside.

# one row per unit · 11-field card schema · hp buckets Under 40 to 300+

Read the pair together and the record's character shows itself. Both units sit on the same Michigan lot, twenty-nine years of depreciation apart: a 7810 at 4,227 hours for $95,900 against an 8320 at 3,516 hours for $129,900. Hours against ask is the whole residual-value game, and the seller notes - always stored inside, one owner - are exactly the condition signals appraisers otherwise pay to collect.

Eleven fields define the default delivery, and every definition was verified against live listing markup during the August 2026 research pass rather than inferred from a column heading. Four identify the machine - year, make, model and category. Three carry the commercial payload - price, hours and drive. Four anchor it to the sale - the seller's comment, location, seller name and photo count.

Two behaviors worth calling out. A missing price survives as its call-for-price marker instead of being silently zeroed, so your averages never swallow the unpriced units. And the seller comment stays verbatim text, which makes it raw material for condition classification rather than a cleaned summary you have to reverse-engineer. Definitions below trace to the captured rows above.

The eleven-column spine ships by default. The layers beneath it fold into additional fields on request because they depend on how deep into each listing the capture goes:

• Serial and stock numbers - the identity layer from detail pages, letting one physical machine deduplicate across marketplaces, auction records and finance filings.
• Model-specific specifications - track size, guidance readiness, transmission, PTO and hydraulic configuration; these populate unevenly by model and flag where absent.
• Auction price results - realized-price context from the Auction Price Data side, carrying the verification note below.
• Upcoming Auctions calendar - forward events as dated rows for pipeline planning.

Name the ones your models need when you ask for the sample and column naming locks against live records at that point.

Geography - the American used-iron market: dealer inventory spread across the lower 48 states, with Canadian units appearing through dealer networks. Search axes include state and distance from a ZIP code, so territory cuts resolve natively without a separate geocoding pass.

Temporal - a current-market picture. Each row is a point-in-time snapshot of one unit, with model years reaching back decades, so the corpus describes today's fleet rather than a historical archive. The auction-side archive depth is not published; treat it as confirm-with-sample before building a realized-price series on it.

Granularity - one row per equipment unit. Tractors carry an extra axis the trade actually argues with: horsepower buckets running from Under 40 HP through 300+ HP, which makes class-level supply reads a filter rather than a derivation.

API, files, or your warehouse. Daily, weekly, or hourly.

Your cadence is decoupled from how fast dealers turn their lots. Take a full fourteen-category pull once to baseline your makes, or keep a warehouse table current so newly listed units and sold-listing exits diff cleanly into yesterday's history - because sold iron leaves the live pool, most teams accumulate their own longitudinal record from the first delivery. Deliveries arrive normalized to the eleven-field dictionary above with call-for-price markers intact and seller comments verbatim, so a Case IH row joins a Kubota row without a cleaning pass. Name your makes, categories and states when you request the sample; the sample ships first regardless.

• Residual value and depreciation modeling - hours against ask across same-model cohorts is the empirical surface lenders, lessors and OEM finance arms fit curves on; see price monitoring.
• Auction-grounded valuation - pair live asks with auction price results to separate advertised value from transaction evidence (heavy equipment auction results).
• Dealer competitive intelligence - count live inventory by state, category and horsepower band to see who is stocking what and where supply is thinning (competitor tracking).
• Collateral valuation - mark equipment loan books against current asks instead of stale book values.
• Aftermarket sizing - the age and horsepower mix of the live fleet is the denominator parts and service planners usually guess at (market sizing).
• ML feature pipelines - typed rows of make, model, year, hours, drive and price train valuation models without a collection layer (ML model training).

Market researchers and consultants (relevance 3/3) get machine-level evidence for US equipment-market sizing instead of trade-press anecdotes - see market researchers use cases. Data scientists and ML engineers (2/3) inherit a flat, typed eleven-field table that trains residual models out of the box - see data scientists use cases. Investors and quant researchers (2/3) read dealer inventory as a cycle indicator, because ask-price dispersion moves ahead of revenue prints - see investors quants use cases. Competitive intelligence product teams (2/3) monitor rival dealer stocking by category and state (competitive intel product teams use cases), sales and growth teams (1/3) rank territories by installed iron (sales growth teams use cases) and developers building data products (1/3) ship price-comparison features on a clean schema (developers builders use cases).

Within agricultural and farm machinery data, this record owns the auction-grounded read on the American used market: live asks across fourteen categories with realized-price context sitting beside them. The neighbors own different jobs. TractorHouse is the larger North American dealer marketplace, with unusually deep specification attributes on every unit. MachineFinder (John Deere) goes deepest on one marquee - roughly 77,500 machines with full serials straight from the dealer network. Mascus and Agriaffaires weight European inventory. TractorData supplies the factory-spec baseline that makes specification fields interpretable, and when the question is fleets rather than markets, FAOSTAT Agricultural Machinery counts tractors in use by country back to 1961. If the question is "what is this class of American iron worth, asked and realized", this is the record that answers it. Browse the rest on the agricultural farm machinery data hub or the best agricultural farm machinery datasets ranking.

Three things worth knowing upfront. First, the ask side dominates: prices are what sellers advertise, and unpriced units survive as call-for-price rows. Auction price results exist as a companion layer but could not be fully verified during the August 2026 research pass, so they confirm with your sample rather than arriving as a settled promise.

Second, the corpus is current-state. Sold units leave the live pool and there is no built-up archive, so longitudinal work accrues from your first delivery onward - repeated snapshots are how turnover gets measured.

Third, totals beyond category level are not published upstream; the tractors count above is the hard number, and we size your exact slice inside the sample request. Everything ships in the eleven-field schema shown above unless you name request-only fields when you ask.