UN Comtrade Plus TradeFlow — Electrical Machinery & Instruments (HS 85/90)

The United Nations' official ledger of who sold what to whom, cut to the two commodity chapters the electronic equipment & instruments industry actually lives in: HS chapter 85 - electrical machinery and equipment, semiconductors, printed circuits, batteries, consumer electronics - and HS chapter 90 - optical, photographic, precision and measuring instruments. UN Comtrade Plus is compiled by the UN Statistics Division from reports submitted by national statistical offices, and every observation lands as a bilateral pair: one reporter country, one partner country, one commodity code, one flow direction, one period, carrying dollar value, net and gross weight, supplementary quantities and explicit estimation flags.

Three properties make this slice earn its keep. Reach: essentially every trading economy appears as a reporter or a partner, so a China-to-USA read and a Vietnam-to-Germany read fall out of the same cube with no re-plumbing. Depth: commodity codes resolve from two-digit chapters down to six-digit lines such as 8542 integrated circuits and 8541 diodes and transistors, with the HS revision stamped per record. Valuation discipline: imports price CIF and exports price FOB, sometimes on the same record, so freight and insurance become a visible spread rather than a buried assumption.

Get a sample of this dataset - name the reporters, partners and commodity lines, and we return real rows shaped exactly like the dictionary below, not screenshots.

Two verified rows from the electronics neighborhood, read straight from the dataset:

period 2022 · reporter 842 USA · flow M import · partner 0 World · cmd 8541 diodes/transistors
cifvalue       17,986,436,849 USD
fobvalue       17,172,864,708 USD
primaryValue   17,986,436,849 USD
netWgt         null          grossWgt 0.0
flags          isReported=false · isAggregate=true

period 2023 · reporter 156 China · flow X export · partner 842 USA · cmd 8542 integrated circuits
fobvalue        2,325,395,161 USD
primaryValue    2,325,395,161 USD
netWgt          1,349,464.296 kg   isNetWgtEstimated=true
legacyEstimationFlag 4             isReported=false · isAggregate=true

Read the pair and three practicalities surface immediately. Both rows are aggregates - isAggregate=true marks chapter-line totals rather than six-digit detail, which arrives in precisely the same column shape under a longer cmdCode, so nothing about your parser changes when you drill down. Valuation rides on direction: the 2022 US import line carries CIF and FOB side by side ($17.99bn against $17.17bn), making the freight-and-insurance wedge a one-column subtraction, while the 2023 China export line quotes FOB alone. And honesty lives in columns, not footnotes - isReported=false says the figure was derived upstream rather than handed over by the national authority, isNetWgtEstimated=true flags that 1,349,464.296 kg specifically, and dividing that weight into the $2.33bn FOB value prices integrated circuits near $1,723 per kilogram before anyone starts arguing about averages.

Sixteen fields anchor the working dictionary, each verified against live records rather than guessed from column names. Identity fields fix who traded what with whom, under which Harmonized System revision, through which customs procedure and transport mode. Measure fields fix money and mass - dollars on the CIF and FOB axes, kilograms on the net-weight axis. The boolean tail separates directly reported figures from derived ones, which is the difference between quoting a number and defending it.

Every row is one bilateral commodity-flow-period observation, so a single year of six-digit detail across all partners runs to millions of rows while chapter-level cuts collapse to thousands.

Additional fields on request: the rest of the dictionary folds underneath - reporter and partner name columns beside the numeric codes, second-partner codes used by re-export flows, the alternative-quantity family with its own unit codes, gross weight, the four estimation flags, aggregation level and leaf-node markers, and the YYYYMMDD period identifier. Column-level naming is confirmed against live records when we prepare your sample - anything outside the core set travels as a labeled add-on, never as a surprise.

Three chips summarize the footprint:

• Geography: Global - reporters and partners covering essentially all trading economies, each row a bilateral pair, so bilateral concentration questions (who actually supplies whom) answer directly instead of through residual math.
• Time frame: Annual series from 1962 onward, with depth varying by classification revision since earlier decades were reported under earlier HS ancestors; monthly series extend over more recent years. Verified live responses returned 2022 and 2023 periods under revision H6.
• Granularity: Bilateral reporter x partner x six-digit HS commodity x flow x period - rollable up to chapter totals and down to leaf commodity lines inside one consistent row shape.

Set against the wider Datadory catalog - average quality score 7.81 across all 1,744 datasets - this slice scores 9/10, carried by universal geographic reach, a six-decade archive and per-record valuation and estimation flags.

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

You choose the channel and the cadence; the field dictionary above travels unchanged through all three. Structured payloads suit event-driven trade monitors, bulk files suit overnight warehouse loads, and tabular extracts land in BI-ready shapes for analysts who never leave the spreadsheet. Cadence changes are a settings conversation, not a re-integration project - and a sample cut to your commodity list comes first either way.

• Investors and quants treat monthly bilateral electronics flows as high-frequency demand signals - import penetration, export momentum and trade-balance screens at resolutions quarterly filings cannot match, with the estimation flags telling them which prints to trust.
• Supply-chain and procurement teams watch component-line flows between named country pairs, so sourcing concentration stops being anecdotal: semiconductor and diode lines shifting between manufacturing geographies show up here months before a supplier review mentions them.
• Competitive-intel and product teams read share shifts inside chapter 85 as evidence in market maps - a partner country's integrated-circuit exports moving while a rival's roadmap deck stays silent.
• Market researchers size electronics markets country by country from official figures, combining import CIF values with domestic production series instead of triangulating from press releases.
• Policy and economics researchers get a six-decade bilateral panel with per-record provenance flags - rare honesty about which observations are reported versus derived - suitable for gravity models and tariff-exposure work without imputation folklore.
• Data scientists and ML engineers load a fixed commodity-country-month vocabulary once and join everywhere: trade rows, industry production data and company fundamentals all key to the same commodity spine.

Investors and quants hold the clearest win: bilateral monthly flows convert into nowcasts that lead quarterly disclosures, and the estimation flags pre-answer the data-quality question every model review asks. Supply-chain teams turn sourcing-risk reviews from interviews into measurements on named commodity lines and country pairs. Developers and builders get a stable grain and a typed field vocabulary that loads once and joins everywhere. The constant across all four: the commodity spine. Everything compounds when trade rows, production statistics and fundamentals speak the same six-digit HS language.

• The electronic equipment & instruments data hub collects every dataset in the industry, this one included.
• Head-to-head: vs Google Patents Public Data on BigQuery - trade values against technology content for the same industry.
• The US-border mirror: Census Foreign Trade Schedule B / HS-based statistics puts American flows at ten-digit detail beside the world cube.
• Vocabulary: Harmonized System codes, CIF and FOB value fields and Comtrade estimation flags explain the layers this dataset stands on.