Patent map¶
A graphical or spatialized analysis of patents in a defined technology field, used to display clusters, relationships, ownership, chronology, whitespace, or possible infringement-relevant overlap under documented search and similarity choices.
Core Idea¶
A patent map is a graphical model of patents in a defined technology space. It can display clusters, citations, assignees, inventors, filing time, classifications, families, or text-derived similarity so analysts can inspect concentrations, relationships, trajectories, and areas with few retrieved documents. The practice is often called patent landscaping. A map is produced, not merely found. A map is produced, not merely found.
Scope of Application¶
Use patent map with technology question, jurisdiction and date scope, search strategy, patent-family treatment, attributes or relation metric, layout, update date, and legal limits stated. Use patent map with technology question, jurisdiction and date scope, search strategy, patent-family treatment, attributes or relation metric, layout, update date, and legal limits stated.
- Competitive intelligence. Compares technology portfolios.
- R&D planning. Finds crowded and sparse areas.
- Prior-art exploration. Navigates relevant documents.
- Portfolio management. Displays ownership and clusters.
- Policy analysis. Tracks technological development.
Clarity¶
Visualization compresses thousands of documents into position and color, improving navigation while hiding claim language and method uncertainty. The closest near miss sets the boundary: A patent landscape report is closest and often includes maps, but the report can contain legal, market, and textual analysis beyond the visual model itself.
Manages Complexity¶
Patent counts and clusters mix strategic filing behavior with invention activity. Families, continuation practice, legal status, and jurisdiction can inflate apparent density. The central overview–claim detail tradeoff is this: A map reveals large-scale organization while legal meaning resides in individual claims. A second visual whitespace–retrieval uncertainty tension matters because Sparse space can indicate opportunity or a search-model blind spot.
Abstract Reasoning¶
Use three linked moves: define the technology and decision question; build and clean the patent corpus; choose attributes or relations appropriate to the question. As a collapse test, the case exits when corpus selection or encoding cannot be explained, or when visual distance is treated as proof of claim overlap or infringement. A fourth check is to encode them with a documented visual method.
Knowledge Transfer¶
Corpus-to-landscape visualization transfers across document analysis, but patent families, claims, classifications, and legal scope delimit patent maps. The nearest stopping boundary is explicit: A patent landscape report is closest and often includes maps, but the report can contain legal, market, and textual analysis beyond the visual model itself. The inclusion test remains: An artifact is a patent map when a defined patent corpus and explicit relation or feature model are encoded graphically to support technology-landscape interpretation. The structure no longer applies when the case exits when corpus selection or encoding cannot be explained, or when visual distance is treated as proof of claim overlap or infringement. The reviewed DAG parent relation is ; it carries the broader structural comparison without erasing the specialist conditions. It is a common synonym and broader practice.
Relationships to Other Abstractions¶
Current abstraction Patent map Domain-specific
Parents (1) — more general patterns this builds on
-
Patent map is a kind of Representation Prime
Patent map is a strict kind of Representation: A patent map is a graphical model of patent visualisation.
Hierarchy path (1) — routes to 1 parentless root
- Patent map → Representation → Abstraction
Neighborhood in Abstraction Space¶
Patent map sits in a sparse region of the domain-specific corpus (66th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Faceted Classification & Metadata (14 abstractions)
Nearest neighbors
- Knowledge organization system — 0.84
- Data Model — 0.84
- Cultural mapping — 0.84
- Data element — 0.84
- Narrative network — 0.84
Computed from structural-signature embeddings · 2026-10-08