Commodity Research Brief v1
First actionable research brief for a Tableau-oriented project comparing coffee, tea, and cocoa, with lighter consideration of yerba mate and matcha only where they genuinely fit. This brief is designed to help the team choose sane source combinations and a clean data model.
1. Comparison dimensions
Price
Primary comparison layer. Use price per kilogram or a directly comparable unit where possible.
Production
Country-year production, harvested area, and related agricultural measures.
Trade
Country-year exports/imports using well-defined commodity codes.
Removed from first-pass core: sales / market size as a primary layer.
2. Top 10 candidate datasets
| Dataset | Best for | Authority | Tableau fit | Notes |
|---|---|---|---|---|
| FAOSTAT Production (QCL) | Production | High | High | Best production backbone for coffee, tea, cocoa. |
| FAOSTAT Trade (TCL) | Trade | High | High | Useful for country-year export quantity/value within FAO system. |
| UN Comtrade | Trade | High | Medium-High | Best detailed trade source; requires HS-code discipline. |
| World Bank Pink Sheet | Price | High | High | Strong cross-commodity monthly price layer in comparable units. |
| FRED coffee price series | Coffee price context | High | High | Strong coffee-only supporting price series. |
| ICO statistics | Coffee validation/context | High | Medium | Excellent support source; some deeper data access may be limited. |
| International Tea Committee | Tea validation/context | High | Low-Medium | Respected, but more publication-oriented and less frictionless to ingest. |
| ICCO statistics | Cocoa validation/context | High | Low-Medium | Great for cocoa-specific support, weaker as a simple first-download source. |
| Our World in Data | Fast prototyping | Medium-High | High | Convenient mirror/convenience source, often FAOSTAT-derived. |
| USDA PSD / FAS | Secondary validation | High | Medium | Potentially useful, but not the simplest first backbone compared with FAOSTAT + World Bank + Comtrade. |
3. Top 5 composition strategies
- Safest: use
World Bank Pink Sheet + FAOSTAT Productionfor price + production. - Best full academic build: use
World Bank Pink Sheet + FAOSTAT + UN Comtradefor price + production + trade. - Keep price separate: monthly price data should stay in its own table, not forced into annual production/trade rows.
- Use a commodity lookup table: keep a small table defining coffee, tea, cocoa concepts and relevant codes.
- Validate with specialist sources: use ICO / ITC / ICCO for context, caveats, and cross-checking rather than building the first fact table directly from their reports.
4. Tableau-readiness comparison
| Source | Authority | Scope | Cleanliness | Tableau-readiness | Reproducibility |
|---|---|---|---|---|---|
| FAOSTAT | 5/5 | 5/5 | 4/5 | 5/5 | 5/5 |
| UN Comtrade | 5/5 | 5/5 | 3/5 | 4/5 | 5/5 |
| World Bank Pink Sheet | 5/5 | 4/5 | 4/5 | 5/5 | 5/5 |
| ICO | 5/5 | 3/5 | 3/5 | 3/5 | 4/5 |
| ITC | 5/5 | 3/5 | 2/5 | 2/5 | 3/5 |
| ICCO | 5/5 | 3/5 | 2/5 | 2/5 | 3/5 |
| OWID | 4/5 | 3/5 | 5/5 | 5/5 | 4/5 |
| USDA PSD | 4/5 | 3/5 | 3/5 | 3/5 | 4/5 |
5. Expert / institutional sources
- ICO — best institutional coffee statistics/context source
- International Tea Committee — best tea-specific institutional statistics/context source
- ICCO — best cocoa-specific institutional statistics/context source
- FAO / FAOSTAT — best official agricultural backbone
- UN Comtrade / UNSD — best official global trade backbone
- World Bank commodity markets / Pink Sheet — strong cross-commodity price context
These are better as trusted institutions than as interchangeable flat-file data sources.
6. Data model recommendation
Recommended Tableau architecture:
prices_monthly— commodity, date/month, unit, valueproduction_country_year— country, year, commodity, production/harvest area/yieldtrade_country_year— reporter, partner(optional), year, commodity, import/export measurecommodity_lookup— commodity definitions and code mappings
Country-year is the preferred architecture for production and trade, but not a red line for price because price data may remain global-monthly.
7. Scope viability options
Coffee + tea + cocoa using price + production
Use World Bank Pink Sheet + FAOSTAT only. Cleanest and most Tableau-ready.
Coffee + tea + cocoa using price + production + trade
Use World Bank Pink Sheet + FAOSTAT + UN Comtrade. Strongest overall research build.
Add yerba / matcha fit analysis
Keep as a sidecar analysis, not the core model. Matcha should remain a tea subcategory if included.
Which source combinations are actually sane to build with?
- Sane: World Bank Pink Sheet + FAOSTAT
- Very sane: World Bank Pink Sheet + FAOSTAT + UN Comtrade
- Sane with validation support: add ICO / ITC / ICCO for context and checking
- Not sane as a first build: PDF-only sources, commercial market reports as backbone, or mixing cocoa with chocolate product data