Uses Claude Sonnet 4 to analyze weather data and generate predictions

Supply risk assessment output

PostgreSQL database for storing all AgriQuant AI data

PyTorch Dataset for time series data

Historical validation of prediction accuracy

Main system coordinator Runs continuous monitoring and generates predictions

Collects weather, vegetation, and supply data for West African cocoa regions. Sources: Ghana Met Agency, NOAA GFS (for Ivory Coast), NASA MOD

Collects and analyzes weather data for Brazilian Arabica coffee regions. Primary data sources: INMET (Brazilian meteorological service),

Analyzes correlations and dependencies across multiple time series: - Weather variables vs OJ prices - OJ vs related commodities (sugar,

Deep learning-based price prediction system Uses LSTM with attention for sophisticated pattern recognition

Sophisticated ensemble prediction model for ag commodity futures prices Combines weather impact, seasonal patterns, and market dynamics

Corn and Wheat weather monitoring across US Corn Belt and Great Plains. Sources: NOAA, USDA Crop Progress reports, US Drought Monitor (USDM),

Satellite-derived grove health metrics

Monitor current grove health conditions Affects vulnerability to weather events

Multi-layer LSTM network for ag commodity futures price prediction Captures temporal dependencies in weather and market patterns

Translate supply disruptions into price impact forecasts Uses: - Supply/demand elasticity - Historical price responses - Cur

Analyzes market microstructure for ag commodity futures - Order flow imbalance - Liquidity metrics - Price impact models - Volume

Tracks trading performance metrics for transparency reporting Calculates accuracy, profitability, risk metrics

Price impact prediction

Process satellite imagery to monitor grove health Data sources: - Sentinel-2 (ESA): Free, 10m resolution, 5-day revisit - Planet

Sugar #11 weather and supply data collection. Monitors Brazil drought, India monsoon, Thailand floods, and policy signals.

Comprehensive supply risk analysis for agricultural crop Combines: - Weather forecasts - Historical patterns - Current grove

Executes monthly token buybacks from trading profits Buys $AGQNT from market and sends to burn address

Distributes USDT airdrops to $AGQNT token holders quarterly 50% of trading profits allocated to USDT distributions

Advanced volatility forecasting for ag commodity futures - GARCH(1,1) model - EWMA (Exponentially Weighted Moving Average) - Regime-s

Monitors ENSO phase using NOAA Climate Prediction Center (CPC) data. Oni Index (Oceanic Nino Index): 3-month running mean of SST anomalies

Weather event types

Central engine that: 1. Runs all 6 commodity collectors in sequence 2. Aggregates signals with correlation adjustments 3. Detects por

Supply risk levels

Download satellite imagery from various sources

Multi-source satellite data processor for vegetation health monitoring. Combines MODIS (free, 250m, 16-day), Sentinel-2 (free, 10m, 5-day),

Monitors and ingests supply/demand reports from all relevant agencies. Detects revision magnitude to assess price impact before market open.