comtrade

comtrade provides clean, tidy access to international trade data from the United Nations Comtrade database directly from R, plus built-in trade analytics (RCA, HHI, trade balance, growth rates, concordance).

What is UN Comtrade?

The United Nations Comtrade database is the largest repository of international merchandise trade statistics in the world. Over 200 countries and territories report their annual and monthly trade data to the UN Statistics Division, which standardises, validates, and publishes it in a consistent format. The database covers goods trade from 1962 to the present and services trade from 2000, with commodity detail down to the 6-digit Harmonized System (HS) level.

Comtrade is the standard data source for bilateral trade flow analysis, gravity models, revealed comparative advantage, trade policy evaluation, and supply chain research. It underpins publications from the World Bank, WTO, UNCTAD, and most academic trade economics research.

How is this different from comtradr?

There is an existing R package called comtradr, maintained by Paul Bochtler through rOpenSci. It wraps the same Comtrade Plus API.

What comtrade adds:

Built-in trade analytics: comtradr is a data pipe only. To compute RCA, HHI, trade balance, or growth rates, you need additional packages (economiccomplexity, concordance) or manual code. comtrade includes ct_rca(), ct_hhi(), ct_balance(), ct_growth(), ct_share(), and ct_compare() out of the box.
Classification concordance: ct_concordance() converts between HS, SITC, and BEC classifications with a built-in lookup table. No separate package needed.
Works without an API key: basic queries use the Comtrade preview endpoint automatically. No registration required for simple trade lookups.
Fewer dependencies: comtrade depends on 3 packages (cli, httr2, tools). comtradr depends on 14 (including lubridate, purrr, poorman, stringr, readr, fs, rappdirs, memoise, cachem, lifecycle, askpass).
CRAN-safe caching: uses tools::R_user_dir() instead of rappdirs.
Returns data.frames: not tibbles. Consistent with base R workflows.

Installation

# From CRAN (when available)
install.packages("comtrade")

# From GitHub
# install.packages("devtools")
devtools::install_github("charlescoverdale/comtrade")

API key (optional but recommended)

Basic queries work without a key. The package automatically falls back to the Comtrade preview endpoint, which returns up to 500 records per call without registration.

For full access (100,000 records per call, commodity descriptions, detailed partner breakdowns), register for a free key at comtradedeveloper.un.org (500 calls/day, no cost):

library(comtrade)
ct_set_key("your-subscription-key")
#> Comtrade API key set for this session.

To make it permanent so you don’t have to set it every session, add it to your .Renviron file. Run file.edit("~/.Renviron") in R, add the line COMTRADE_API_KEY=your-key, save, and restart R. The package will pick it up automatically from then on.

Examples

Bilateral trade flows

library(comtrade)

# UK total exports to the world, 2023
uk <- ct_trade("GBR", flow = "X", year = 2023)
head(uk)
#>   reporter reporter_desc partner partner_desc flow flow_desc commodity_code
#>        826           GBR       0        World    X   Export          TOTAL
#>   commodity_desc year period trade_value_usd net_weight_kg quantity quantity_unit
#>            Total 2023   2023   4.685e+11             0        0          <NA>

Top export products

# Australia's top 5 exports by HS chapter
ct_top_products("AUS", flow = "X", year = 2023, n = 5)
#>   commodity_code      commodity_desc        value share_pct rank
#>               26  Ores, slag, ash    1.256e+11     33.21    1
#>               27  Mineral fuels      1.081e+11     28.60    2
#>               71  Precious stones     2.014e+10      5.33    3
#>               02  Meat             1.043e+10      2.76    4
#>               10  Cereals          5.874e+09      1.55    5

Trade balance

# US trade balance with China, 2020-2023
ct_balance("USA", partner = "CHN", year = 2020:2023)
#>   partner partner_desc year      exports      imports      balance
#>       156        China 2020  1.246e+11   4.352e+11  -3.106e+11
#>       156        China 2021  1.511e+11   5.063e+11  -3.552e+11
#>       156        China 2022  1.538e+11   5.368e+11  -3.830e+11
#>       156        China 2023  1.478e+11   4.272e+11  -2.794e+11

Revealed comparative advantage

# Which products does Australia have a comparative advantage in?
rca <- ct_rca("AUS", year = 2023, level = 2)
head(rca[rca$has_advantage, ], 5)
#>   commodity_code commodity_desc  reporter_value  world_value  rca has_advantage
#>               26 Ores, slag      1.256e+11     3.95e+11   18.42          TRUE
#>               02 Meat            1.043e+10     1.29e+11    4.69          TRUE
#>               10 Cereals         5.874e+09     8.05e+10    4.23          TRUE
#>               51 Wool            2.817e+09     1.18e+10   13.84          TRUE
#>               04 Dairy, eggs     3.122e+09     4.67e+10    3.88          TRUE

Trade concentration (HHI)

# How concentrated are Australia's export destinations?
ct_hhi("AUS", flow = "X", year = 2023, by = "partner")
#>   year  hhi concentration n_items     top_item top_share_pct
#>   2023 1847      moderate     196        China         32.71

Trade growth

# UK export growth, 2018-2023
ct_growth("GBR", flow = "X", years = 2018:2023)
#>   year        value growth_yoy growth_cumulative index_100
#>   2018  4.871e+11         NA              0.00    100.00
#>   2019  4.693e+11      -3.65             -3.65     96.35
#>   2020  3.705e+11     -21.04            -23.93     76.07
#>   2021  4.037e+11       8.96            -17.12     82.88
#>   2022  5.058e+11      25.29              3.84    103.84
#>   2023  4.685e+11      -7.37             -3.82     96.18

Classification concordance

# What SITC section does HS chapter 27 (mineral fuels) map to?
ct_concordance("27", from = "HS", to = "SITC")
#>   from_code          from_desc to_code       to_desc
#>          27 Mineral fuels, oils       3 Mineral fuels

Functions

Data retrieval

Function	Purpose
`ct_trade()`	Bilateral goods trade flows (200+ countries, 1962-present)
`ct_services()`	Services trade via EBOPS classification (2000-present)
`ct_reporters()`	List available reporter countries
`ct_commodities()`	Search HS commodity codes
`ct_available()`	Check data availability for a country

Trade analytics

Function	Purpose
`ct_balance()`	Trade balance (exports minus imports) by partner
`ct_top_products()`	Top N products by trade value with shares
`ct_top_partners()`	Top N trading partners by value with shares
`ct_rca()`	Revealed Comparative Advantage (Balassa index)
`ct_hhi()`	Herfindahl-Hirschman concentration index
`ct_growth()`	Year-on-year and cumulative trade growth
`ct_share()`	Country’s share of world trade
`ct_compare()`	Compare multiple countries’ trade

Utilities

Function	Purpose
`ct_concordance()`	Convert between HS, SITC, and BEC classifications
`ct_set_key()`	Store API key
`ct_cache_clear()`	Clear cached responses

Caching

Results are cached locally after the first download so repeated calls are instant and don’t touch the API. The cache directory defaults to tools::R_user_dir("comtrade", "cache") and can be overridden with options(comtrade.cache_dir = "/your/path"). Cache entries expire after 24 hours for trade data and 7 days for reference tables.

Data source

All data comes from the United Nations Comtrade database via the Comtrade Plus API. This package is not affiliated with or endorsed by the United Nations.

Issues

Found a bug or have a feature request? Open an issue at github.com/charlescoverdale/comtrade/issues.

Keywords

trade, international trade, comtrade, UN, bilateral, exports, imports, HS, SITC, BEC, RCA, HHI, trade balance, gravity model, comparative advantage

License

MIT

mirror server hosted at Truenetwork, Russian Federation.