For Data Scientists & ML Engineers · Trading Companies Distributors
Trading Companies & Distributors Data for Data Scientists
Trading Companies Distributors data for data scientists: 15 datasets on one shelf. Every one delivered as API, files, or warehouse rows.
financial time series api for backtesting · alternative data for quantitative research · where to get training data for trading companies distributors models
API, files, or your warehouse. Daily, weekly, or hourly.
Which trading companies & distributors datasets should data scientists pull first?
We rank for pipeline fit: three relevance-3 records lead, nine relevance-2 records follow and three relevance-1 records trail, ordered inside each band by quality score.
How do you build a wholesale demand model from these sources?
Nowcasting leads. Short-term business statistics — Wholesale & Retail Trade (NACE G46) is the top-ranked series: one monthly feed back to January 1991, downloadable as JSON, TSV, CSV or XLSX, pairing directly with Table 20-10-0074-01, a single CSV zip covering 1993-2026 at six-digit NAICS depth.
Classic-ML work keeps its own lane: the UCI Wholesale Customers Dataset's 440-customer spend matrix remains the standard fixture for regression-testing clustering and segmentation pipelines, shipped statically as CSV and ZIP. Entity linkage starts from Companies House monthly snapshots filtered to SIC 46 — roughly 5-6 million live UK companies for survival modelling.
How fresh is each source, and does cadence limit the modelling?
Cadence splits along the relevance bands. The 4 daily refreshers — data.europa.eu, data.gov, ImportGenius and Panjiva — are discovery catalogs and paywalled shipment feeds, not modelling staples, and Alibaba.com churns listing pages rather than publishing series.
For scale, 22.6% of the full 1744-dataset catalog.8% refreshes at least weekly; this panel runs 27% daily.
What is the bottom line for data scientists?
Build in evidence order: wire the NACE G46 series and the Canadian sales table together as the nowcasting core, layer the annual survey and GDP-by-industry aggregates underneath for structural features, and keep the UCI matrix for teaching rather than production signal. Six of 15 records here score 8-plus on quality, below the 62.8% catalog-wide share, because much of this slice is discovery metadata rather than deep series. Comparable slices sit on our all data-scientists resources page.
Straight answers
Is there a financial time series API for backtesting distributor demand?
Cadence is monthly to annual, so backtest on low-frequency features.
Where do I get training data for wholesale demand forecasting models?
Start with the two monthly series — NACE G46 short-term statistics and Table 20-10-0074-01 — then enrich them with the annual survey's revenue, COGS and margin series and GDP-by-industry KLEMS components back to 1947. That yields decades of labelled history for feature stores. WTO country-pair panels extend the same models with trade-flow features, subject to its commercial-use permission.
Rows before rollout
Sample rows from any shelf entry — the field dictionary and coverage notes ride along. If the shelf misses what you need, say so; sourcing requests are half our job.
Talk to us