DATA METHODOLOGY & AI PROCESSING FRAMEWORK

Effective Date: January 20, 2026

1. PRINCIPLES OF DATA SOURCING
1.1. Strict Public Domain Sourcing. The Platform aggregates trade, logistics, and chemical volume data exclusively from official, publicly available government databases, international open-source registries, and published federal statistical records (such as USITC, EPA, Eurostat, and global customs books).
1.2. No Proprietary Data Intake. The Platform does not collect, request, scrape, or process private, non-public, commercially sensitive, or proprietary transaction data from active market participants.
1.3. Exclusion of Individual Identifiers. All raw data utilized by our AI models is strictly anonymized and aggregated at the macro, regional, or industry-wide level. The Platform cannot and does not trace data back to specific private corporations, individual commercial invoices, or confidential company margins.
2. DATA PROCESSING AND HISTORICAL ANALYSIS
2.1. Retrospective/Historical Focus. The algorithms process historical, backward-looking datasets. The input data typically reflects market realities with a standard reporting lag inherent to official government updates (ranging from several weeks to months).
2.2. Statistical Aggregation and Outlier Removal. To ensure data integrity, the Platform applies standard mathematical moving averages, variance filters, and statistical anomaly detection rules. Any isolated or unverified data points are programmatically discarded to prevent artificial market skewing.
3. AI INSIGHT GENERATION AND NON-BINDING PREDICTIONS
3.1. Algorithmic Probability Modeling. The Platform’s AI models analyze historical correlations between supply, demand, regulatory shifts, and trade volumes to generate analytical forecasts. These forecasts represent statistical probabilities and mathematical trends, not factual certainties.
3.2. Low-Risk General Purpose AI. The AI functions strictly as a diagnostic and exploratory data tool under global regulatory frameworks (including the EU AI Act). It does not execute automated trades, does not interact via API with live trading books, and does not possess direct decision-making capabilities for any third-party business operations.
3.3. No Market Mandates. The outputs generated by the AI model are entirely non-binding. The Platform provides market visualization and historical pattern recognition to assist independent human research. It does not issue pricing directives, commercial targets, or instructions on market behavior.