How to use named entity recognition in ATLAS.ti

Key takeaways

  • Named Entity Recognition (NER) uses machine learning to identify and code named entities in textual data.
  • NER can identify entities such as people, locations, organizations, and miscellaneous named entities.
  • The feature is currently available in ATLAS.ti Desktop (Windows and Mac) and is not currently supported in ATLAS.ti Web.
  • Researchers can review detected entities before applying codes.
  • NER can analyze documents, document groups, or quotations associated with selected codes.

Who this article is for

This article is for ATLAS.ti Desktop users who want to automatically identify and code named entities such as people, organizations, and locations in textual data.

What is Named Entity Recognition?

Named Entity Recognition (NER) is an AI-assisted analysis tool that automatically identifies named entities within text.

Named entities typically include:

  • people
  • locations
  • organizations
  • miscellaneous entities

The tool analyzes selected data, identifies named entities, and proposes corresponding codes that can be applied during analysis.

NER can help researchers:

  • identify frequently mentioned people or organizations
  • analyze stakeholder relationships
  • explore geographic references
  • support content analysis workflows
  • accelerate coding of structured information
  • identify recurring named entities across datasets

Availability of Named Entity Recognition

Named Entity Recognition is currently available in:

  • ATLAS.ti Windows
  • ATLAS.ti Mac

Named Entity Recognition is not currently available in ATLAS.ti Web.

How Named Entity Recognition works

The NER tool analyzes selected textual data and identifies entities based on machine learning language models.

The tool can detect entities in categories such as:

  • Person
  • Location
  • Organization
  • Miscellaneous

Researchers can:

  • review detected entities
  • modify entity categories if needed
  • select which entities to code
  • apply proposed codes individually or in bulk

How to use Named Entity Recognition in ATLAS.ti Desktop

Named Entity Recognition can analyze:

  • documents
  • document groups
  • quotations associated with selected codes

Step 1: Open Named Entity Recognition

  • In ATLAS.ti Windows
  1. Open your project.
  2. Go to the Search & Code tab.
  3. Select Named Entity Recognition.

  • In ATLAS.ti Mac
  1. Open your project.
  2. Click Analysis.
  3. Select Named Entity Recognition.

Step 2: Select the data to analyze

  • In ATLAS.ti Windows and Mac
  1. Select the documents or document groups you want to analyze.
  2. Alternatively, click Codes and select codes if you want to analyze quotations associated with specific codes.

Step 3: Select the analysis unit

Choose the level of text that ATLAS.ti should analyze.

  • In ATLAS.ti Windows

Options include:

  • Paragraphs
  • Sentences

  • In ATLAS.ti Mac

Options include:

  • Paragraphs
  • Sentences
  • Words
  1. Select the preferred analysis unit.
  2. Continue to the next step.

Smaller units may identify more precise entities, while larger units provide additional context.

Step 4: Select entity categories

Choose which types of entities ATLAS.ti should identify.

Available categories include:

  • Person
  • Location
  • Organization
  • Miscellaneous
  • In ATLAS.ti Windows and Mac
  1. Mark the entity categories you want to analyze
  2. Continue.

Step 5: Manage language models if needed

ATLAS.ti uses language models to identify named entities.

  • In ATLAS.ti Windows and Mac
  1. Click Manage Models if you want to change the language model.
  2. Install or select the desired model.

Step 6: Review identified entities

After processing the data, ATLAS.ti displays the entities it has identified.

  • In ATLAS.ti Windows and Mac
  1. Review the list of detected entities.
  2. Select the entities you want to code.
  3. Change the entity category if necessary.

Step 7: Review the search results

ATLAS.ti generates quotations associated with the selected entities.

  • In ATLAS.ti Windows and Mac
  1. Review the quotations in the results page.
  2. Examine any proposed codes.

Proposed codes appear as grayed-out text with dotted outlines, while existing codes appear as solid code labels.

Step 8: Apply codes

You can apply codes individually, by entity, or across all results.

Apply individual codes

  • In ATLAS.ti Windows and Mac
  1. Review a quotation.
  2. Click the plus (+) icon next to the proposed code.

Apply all proposed codes

  • In ATLAS.ti Windows and Mac
  1. Click Apply Proposed Code(s).
  2. Choose whether to apply all results or all results for a specific entity.

Apply custom codes

  • In ATLAS.ti Windows and Mac
  1. Select the desired quotations.
  2. Click Code icon
  3. Choose existing codes or create new codes.

When to use Named Entity Recognition

Named Entity Recognition can be useful when you want to:

  • identify frequently mentioned people
  • analyze organizations referenced in documents
  • explore geographic references
  • identify key stakeholders
  • analyze media or policy documents
  • study social networks and relationships
  • support content analysis projects

Common use cases include:

  • interview analysis
  • news and media analysis
  • policy research
  • organizational research
  • stakeholder analysis
  • public discourse studies

Best practices for Named Entity Recognition

To improve analysis quality:

  • review detected entities before coding
  • verify entity categories are correct
  • check for duplicate or inconsistent entity names
  • use custom codes where additional detail is needed
  • combine NER with manual coding and interpretation

NER works best when used as a starting point for further qualitative analysis rather than as a replacement for researcher interpretation.

Common issues and mistakes

  • Expecting Named Entity Recognition in ATLAS.ti Web
    • Named Entity Recognition is currently available only in ATLAS.ti Desktop.
  • Applying all proposed codes without review
    • Always review entities and quotations before applying codes across the entire dataset.
  • Selecting too many entity categories
    • Selecting all categories may generate a large number of results that require additional review.
  • Using the wrong analysis unit
    • Word-level analysis may generate highly specific results, while paragraph-level analysis may provide more context. Choose the option that best matches your research goals.
  • Assuming all detected entities are correct
    • Machine learning models may occasionally misclassify entities or assign the wrong category.

When to contact support

Contact ATLAS.ti Support if:

  • Named Entity Recognition does not open
  • entity results fail to generate
  • language models cannot be installed or updated
  • entity categories do not appear correctly
  • quotations are missing from the results
  • proposed codes cannot be applied
  • the tool behaves differently from the documentation

When contacting support, include:

  • your platform: Windows or Mac
  • your ATLAS.ti version
  • screenshots or error messages
  • the type of data you are analyzing
  • the selected analysis unit
  • a description of your workflow
Did this answer your question? Thanks for the feedback There was a problem submitting your feedback. Please try again later.