TRANSPARENT INNOVATION INTELLIGENCE

Data Methodology & Provenance Architecture

IDiyas ingests, structures, and links patent, trademark, and prosecution data across millions of records. This document outlines our authoritative data pipelines, entity resolution models, update cadence, and analytical scoring frameworks.

1. Primary Data Sources & Coverage

IDiyas aggregates public government and intellectual property registry data without synthetic bias:

United States Patent and Trademark Office (USPTO): Full-text granted utility, design, and plant patents from 1976 to the present; published patent applications (PGPubs) from 2001 to the present; trademark registers; and Patent Center prosecution records.
European Patent Office (EPO): Granted patents, publication numbers, and patent applications.
Canadian Intellectual Property Office (CIPO): Canadian granted patents and applications.
Cooperative Patent Classification (CPC): Hierarchical technology classifications administered jointly by the USPTO and EPO.

2. Update Frequency & Synchronization

The USPTO officially issues granted patents weekly every Tuesday at 00:00 EST. IDiyas automated ingestion pipelines execute within hours of the official data release:

Parse and index newly granted patents, citations, assignees, and inventors.
Recompute backward and forward citation graphs across all previously granted patents.
Update ranking tables, geographic innovation indices, and corporate technology portfolios.
Trigger automated IndexNow pings for search discovery on Bing, Perplexity, and Yandex.

3. Entity Disambiguation & Graph Normalization

Raw patent data frequently suffers from spelling irregularities, corporate subsidiary divergence, and inventor name collisions. IDiyas employs graph-based entity normalization:

Inventor Disambiguation: Disambiguates homonyms and clusters aliases using co-inventor relationships, corporate assignees, geographic movement patterns, and CPC subclass consistency.
Corporate Assignee Normalization: Maps variations (e.g., "Qualcomm Inc.", "Qualcomm Incorporated") to canonical parent entities while preserving original assignment filings.
Law Firm & Practitioner Mapping: Normalizes attorney records and correlates firm prosecution portfolios across USPTO art units.

4. Analytical Scoring & Metrics

IDiyas introduces quantitative metrics designed to evaluate technology defensibility and prolific innovation:

Forward Citations (Non-Self): Isolates external citations made by subsequent patents, removing self-citation inflation to reflect genuine market impact.
pH-Index: An h-index adaptation for patent portfolios, denoting that an inventor has h patents each cited at least h times.
Claim Defensibility: Analyzes claim length, independent claim breath, and prior art references to assess vulnerability to invalidation.

5. Technical Methodology Papers

For in-depth mathematical formulations and clustering methodologies, access our official whitepapers:

Look-Alike Inventors Disambiguation Methodology (PDF) Look-Alike Attorneys & Firm Analytics Methodology (PDF) Metropolitan Innovation & City Ranking Methodology (PDF)