Review project information, add or remove data, manage language-processing settings and apply changes through a rebuild. Project Settings contains the controls that affect a specific text analysis project. Some changes take effect immediately, while changes to the project’s data or language model are prepared for the next build.
Open Project Settings
Open a text analysis project and select Project Settings from the bar navigation.
The text analysis project overview displays information such as:
Text analysis project name and description
Creation date and creator
Workspace
Number of records
Text analysis project language
Science version
Total record count
Date of the most recent build
The options available on this page may depend on your permissions.
Edit the text analysis project name or description
Use the edit icons beside the name or description to update how the text analysis project is identified. A useful description should explain what the data represents, the purpose of the analysis and any relevant scope such as a period, source, audience, or business question. Changing the name or description does not rebuild the text analysis project or alter its analysis.
Understand the current and next build
Some Project Settings changes affect how Journey Insights Hub processes the data. These changes appear under Next build and do not affect the current analysis until the project is rebuilt.
Review science assertions for the next build and trigger a rebuild in this section.
This section contains:
Science assertion summary
Rebuild button
Estimated rebuild time
The option to rebuild is available only when the text analysis project has changes that require a new build, such as:
Pending science assertions
A pending science update
A pending language change
When there are no pending changes, Next build displays No changes. After making a change that requires processing, select Rebuild to apply it. A rebuild can change the concepts, prevalence rankings, relationships and other results shown throughout the text analysis project.
Change the text analysis project language
The selected language determines how Journey Insights Hub processes the text analysis project text.
To prepare a language change:
Find the Language section
Select the appropriate language
Select Update at next build
Review the pending change and rebuild the text analysis project when ready
Important: Only change the language when the current selection is incorrect. A language change can substantially alter the concepts, matches and relationships because the text must be processed using a different language model.
For multilingual data, separate the records by language and create a different text analysis project for each language.
Configure collocation settings
A collocation is a group of words that Journey Insights Hub recognizes and analyzes as a single concept. Project Settings controls the maximum number of words that may be combined automatically into a collocation. The default maximum is five concepts. Select Edit to change the displayed limit.
A higher limit permits longer expressions to be identified as concepts, but it may also produce phrases that are too specific to be broadly useful. Use the default unless the text analysis project consistently contains meaningful longer expressions that need to be analyzed together. This setting applies during a build.
Manage Science assertions
Science assertions provide project-specific instructions about how Journey Insights Hub should recognize or treat particular words and phrases.
Assertions affect the project’s underlying language model rather than only changing how a concept is displayed. Use them when the language in the dataset requires a consistent interpretation that the project does not identify automatically.
Science assertions may introduce bias when used without careful review.
When to use a Science assertion
Science assertions are useful when:
A frequent but unhelpful term should not influence the analysis
Multiple terms should be interpreted as the same concept
A meaningful term is being treated as a stop word
A multi-word expression should be recognized as one concept
For ordinary analysis themes, use concept lists or advanced concepts instead. Science assertions change how the text analysis project itself is modeled and should be reserved for changes to the underlying language processing.
Assertion types
Ignore turns selected concepts into stop words. Ignored concepts no longer contribute to prevalence ranking or concept-to-concept associations. Use Ignore for terms that occur frequently but add little analytical value in the context of the text analysis project.
Ignoring a term affects its role throughout the text analysis project. Do not use it simply because the term is unimportant to one specific analysis question; a filter or concept selection may be more appropriate in that situation.
Merge instructs Journey Insights Hub to treat selected terms as a single concept. Use Merge when different terms should have the same meaning within the text analysis project, such as a standard name and an abbreviation.
Merge is different from creating a theme or advanced concept. A concept list organizes related concepts for analysis, while Merge changes how terms are represented in the text analysis project model.
Notice ensures that an important text analysis project term is recognized when it might otherwise be interpreted as a stop word or as a common expression with little analytical meaning. Use Notice when a word has a specific, meaningful use in your dataset that differs from its usual grammatical or general-language use.
Force Collocate instructs Journey Insights Hub to recognize a multi-word expression as one concept even when it does not appear frequently enough to be detected automatically.
Use it when the complete phrase has a distinct meaning that would be lost if its words were analyzed separately. Before forcing a collocation, review matching records to confirm that the phrase is used consistently and represents a meaningful idea across the text analysis project.
Edit an assertion
The Science assertions section displays each assertion type with two columns:
Current build shows the assertions presently applied to the text analysis project
Next build shows what will be applied after the next rebuild
To prepare a change:
Find the appropriate assertion type
Select Edit
Add, modify, or remove the relevant terms
Save the change
Review the resulting Next build column
Select Rebuild when you are ready to apply it
Until the text analysis project is rebuilt, the existing analysis continues to use the Current build settings.
Evaluate the effect of an assertion
After rebuilding:
Search for the affected terms
Review their matches in Documents
Check how related concepts and prevalence rankings have changed
Examine the relevant analysis views for unexpected effects
Compare the results with the analytical purpose of the assertion
Assertions can influence more than the term being edited because the text analysis project model learns relationships across the complete dataset. Make one purposeful group of changes at a time so that their effects are easier to evaluate.
Remove records
Use Record Management when specific records need to be removed from the text analysis project.
Select Select Records
Identify the records you want to remove
Carefully review your selection before confirming
Rebuild the text analysis project
Removing records changes the population being analyzed and can affect every text analysis project result. Preserve a copy of necessary source data before deleting records.
Add data
Select Add data to add records to an existing text analysis project. The new records are prepared for the next build.
Before adding data, confirm that:
Each row represents one record
At least one column contains open-ended text
New columns align with the existing text analysis project fields
Dates, scores, and other metadata use consistent formats
The added data belongs within the text analysis project’s analytical scope
Data from multiple datasets can be added even when the files contain different text and metadata fields. Fields that are not present in every source remain blank for the records where they do not apply. Review any warnings about fields that do not match the existing project before continuing. The project must be rebuilt before the new records are incorporated into the analysis.
Copy a text analysis project
Select Copy project when you need a separate text analysis project based on an existing one.
This can be useful for testing changes without altering the original analysis. Review the copy after it is created to confirm which data, settings and analysis resources were carried over before using it as a replacement for the original text analysis project.
Delete a project
Select Delete project to permanently remove the text analysis project.
Deleting a text analysis project is different from removing selected records. It removes the complete project and its analysis resources. Confirm that the project is no longer needed and that any required data or results have been downloaded before continuing.