Entity resolution
Entity resolution is the process of determining which records or mentions refer to the same real-world entity despite differences in their representations.
The Well-Grounded Data Analyst
Entity resolution is the process of determining which records refer to the same real-world entity, including through deduplication and record linkage.
- How it is used
- Builds a customer-level data model from purchases, CRM, and customer records for reliable customer counting.
- How deep it goes
- Works through exact deduplication and library-based fuzzy linkage, while deferring deeper algorithm implementation.
- Examples
- Uses sales, CRM, customer-database, guest-checkout, and varying company-name records to test duplicate decisions.
- What you can do after
- Apply it via pandas and the Python Record Linkage Toolkit to identify, compare, and flag likely duplicate customer records.
Read how The Well-Grounded Data Analyst explains entity resolution →
First-Party Data Activation
The process of reconciling and unifying records that refer to the same real-world entity, such as a person, household, business, or account.
- How it is used
- Unifies fragmented customer, household, or company records so marketing activity can be associated with the intended entity.
- How deep it goes
- Defines the task, compares matching strategies and errors, and discusses platform trade-offs without providing an implementation walkthrough.
- Examples
- Uses a multichannel shopper’s conflicting identifiers, a university’s false-positive match, and split company accounts to illustrate matching risks.
- What you can do after
- Evaluate/understand matching approaches by weighing transparency, reach, false matches, and the operational cost of mistaken merges.
Read how First-Party Data Activation explains entity resolution →
Knowledge Graphs and LLMs in Action
Entity resolution is the process of determining which records or mentions refer to the same real-world entity.
- How it is used
- Integrates heterogeneous records into coherent, queryable knowledge graphs by linking mentions across documents.
- How deep it goes
- Defines resolution and outlines a string- and graph-similarity baseline, but stops short of a full algorithmic treatment.
- Examples
- Uses abbreviated person names, shared employers, research topics, and clustered occupation embeddings as concrete resolution cases.
- What you can do after
- Build it from scratch using candidate string matches, graph context, domain thresholds, and unsupervised clustering.
Read how Knowledge Graphs and LLMs in Action explains entity resolution →
AI-Powered Search
Entity resolution is the process of determining when different representations refer to the same real-world entity.
- How it is used
- Connects related unstructured documents when references to an entity vary across names, pronouns, media, or phrases.
- How deep it goes
- Explains fuzzy foreign keys through concrete references and notes that natural language processing is required beyond strict token matching.
- Examples
- Uses “Trey Grainger,” “Trey,” “he,” “his,” an image, and a video as references to one entity.
- What you can do after
- Apply fuzzy entity references to join information across unstructured documents.
Essential GraphRAG
Entity resolution is the process of identifying and merging different representations of the same real-world entity within a dataset or knowledge graph.
- How it is used
- Cleans knowledge graphs by merging duplicate entity representations before precise querying and inference.
- How deep it goes
- Defines the task, illustrates naming variation, and names methods while deferring implementation to domain-specific rules and thresholds.
- Examples
- Uses three variants of UTI Asset Management Company and a figure showing potential duplicate nodes.
- What you can do after
- Apply it via string matching, clustering, machine learning, or domain-specific rules to unify duplicate references.
Related concepts
- customer data platformused by · A customer data platform is a platform that unifies customer data from multiple sources into persistent customer profiles for activation and analysis.
- user profilepart of · A user profile is a collection of information associated with a user or account, such as identifying attributes, preferences, settings, or activity.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.