Data Sources¶

agrobr integrates data from 40 agricultural data sources. All sources support return_meta=True for full traceability.

Overview¶

Source	Type	Update	Coverage
CEPEA/ESALQ	Prices	Daily	Agricultural commodities
CONAB	Crops, costs, historical series, weekly progress, CEASA/PROHORT	Monthly/Weekly/Daily	National production
IBGE/SIDRA	Statistics	Annual/Monthly/Quarterly	Official data (PAM, LSPA, PPM, Slaughter, PEVS, Milk, GDP, Census)
NASA POWER	Climatology	Daily	Global, 0.5 degree grid
BCB/SICOR	Rural credit, time series, exchange rates, forecasts	Monthly	Crop/state (+ BigQuery)
ComexStat	Exports	Weekly	NCM/state
ANDA	Fertilizers	Monthly	State/month
ABIOVE	Soybean complex exports	Monthly	Volume/revenue
ANEC	Weekly shipments by port	Weekly	19 ports, 6 products
USDA PSD	International supply/demand	Monthly	Global commodities
IMEA	MT quotes and indicators	Daily	Mato Grosso
DERAL	PR crop condition	Weekly	Paraná
INMET	Meteorology	Daily	600+ stations (data requires token)
Notícias Agrícolas	Quotes (CEPEA fallback)	Daily	Commodities
Queimadas/INPE	Fire hotspots	Daily	6 biomes, 13 satellites
Deforestation PRODES/DETER	Deforestation + alerts	Annual/Daily	Amazônia, Cerrado, Pantanal
MapBiomas	Land cover and use	Annual	Municipalities (1985-present)
B3 Agricultural Futures	Daily settlements + open interest	Daily	7 agricultural contracts
UN Comtrade	Bilateral trade + trade mirror	Monthly/Annual	~200 countries, HS codes
ANTAQ	Port cargo movement	Annual	Brazilian ports, 2010+
ANP Diesel	Resale prices + diesel volumes	Weekly/Monthly	States, municipalities, 2013+
ANTT Pedagio	Vehicle traffic at toll plazas	Monthly	200+ plazas, 2010+
MAPA PSR	Rural insurance policies and claims	Annual	27 states, 2006+
SICAR	Rural Environmental Registry (CAR)	Continuous	27 states, 7.4M+ properties
ZARC	Agricultural Climate Risk Zoning	Weekly	40+ crops, all municipalities
Agrofit/MAPA	Registered pesticides	Continuous	~8K formulated products, ~267K authorizations
FUNAI Indigenous Lands	Indigenous lands (WFS geo)	Continuous	~740 TIs, all states
ICMBio Federal Conservation Units	Federal conservation units (WFS geo)	Continuous	344 federal UCs
INCRA Quilombola Territories	Quilombola territories (WFS geo)	Continuous	~426 territories
Land Registry/INCRA	Certified parcels + settlements (shapefile ZIP)	Continuous	SIGEF (15 states) + SNCI (10 states) + settlements Brazil
IBAMA Environmental Embargoes	Environmental embargoes (SIFISC CSV + WKT)	Monthly	~114K embargoes
MapBiomas Alerta	Deforestation alerts (GraphQL)	Weekly	National
Lista Suja	Forced-labor registry (PDF)	Semiannual	National
ANA/SNIRH	Hydrography, irrigation, water availability (ArcGIS REST)	Variable	National
SFB	Public forests, concessions, IFN (ArcGIS REST)	Annual	National
RNC/CultivarWeb	Registered/protected cultivars	Continuous	~37K registered, ~5K protected
EMBRAPA Solos	Soil profiles and soil map	Continuous	34K profiles, 2.8K polygons
Fundação Rio Verde	MT soybean cultivar trials	Annual	~97 cultivars x 4 seasons
CFTC COT	Fund positioning in agricultural futures	Weekly	12 Chicago/NY contracts, 2006+
UNICA	Center-South sugar/ethanol crushing and production	Biweekly	Current crop year + history 1980-2021

Provenance and Traceability¶

All information returned by agrobr can be traced back to its origin. Use the return_meta=True parameter to obtain full provenance metadata.

import asyncio
from agrobr import cepea

async def main():
    # Basic usage (unchanged)
    df = await cepea.indicador('soja')

    # With provenance metadata
    df, meta = await cepea.indicador('soja', return_meta=True)

    print(f"Source: {meta.source}")
    print(f"URL: {meta.source_url}")
    print(f"Fetched at: {meta.fetched_at}")
    print(f"From cache: {meta.from_cache}")
    print(f"Records: {meta.records_count}")

asyncio.run(main())

MetaInfo Structure¶

The MetaInfo object contains the following information:

Field	Type	Description
`source`	str	Source name (cepea, conab, ibge)
`source_url`	str	Exact URL accessed
`source_method`	str	Access method (httpx, cache)
`fetched_at`	datetime	Collection timestamp
`from_cache`	bool	Whether it came from the local cache
`cache_key`	str	Cache key
`cache_expires_at`	datetime	When the cache expires
`records_count`	int	Number of records
`columns`	list	Returned columns
`fetch_duration_ms`	int	Fetch time in ms
`parse_duration_ms`	int	Parsing time in ms
`agrobr_version`	str	agrobr version
`parser_version`	int	Parser version used

Integrity Verification¶

MetaInfo lets you verify data integrity:

# Check that the DataFrame was not altered
is_valid = meta.verify_hash(df)

# Export metadata for auditing
meta_json = meta.to_json()
meta_dict = meta.to_dict()

Diagnostics¶

Use the doctor command to check system health:

agrobr doctor

Returns: - Source connectivity status - Cache statistics - Latest collections - Current configuration