24 Sep 2026

Replicating Oracle Analytics Server Narrative Views in Power BI

Narrative views in Oracle Analytics Server (OAS) turn dashboard data into dynamic, data-driven text, contextual explanations, conditional alerts, and formatted commentary sitting alongside the visualizations they describe. Power BI has no single, direct equivalent, so replicating this functionality during a migration means choosing the right technical approach for each narrative's complexity, not applying one universal fix.

Quick Summary

OAS narrative views, dynamic text, conditional formatting, and data-driven alerts are rebuilt in Power BI using three main approaches: DAX measures that return text, Power BI's native Smart Narrative visual and Copilot-generated summaries, or R/Python script visuals for advanced cases. The right method depends on each narrative's complexity. Security, performance, and validation need the same rigor as any other migrated report element, not an afterthought.

Key Takeaways

OAS narrative views have no direct Power BI equivalent. Each one is rebuilt using DAX text measures, native narrative visuals, or script visuals, chosen by complexity.
A third-party visual mentioned in older guidance, "Narrative Science," was removed from the Power BI marketplace years ago and is no longer available. Power BI's native Smart Narrative visual and Copilot summaries are the current, supported path.
Data modeling comes first. Calculated columns, text-generating measures, and correctly optimized relationships are the foundation every narrative technique depends on.
Security needs dedicated attention. Narrative text can leak sensitive detail even when the underlying visual respects row-level security correctly; this needs its own testing pass.
Migration is also a modernization opportunity. Power BI's AI and natural-language capabilities can genuinely exceed what OAS narrative views offered originally, not just replicate them.

What Are OAS Narrative Views, and Why They're Hard to Replicate

OAS narrative views generate dynamic, data-driven text: values that change based on the underlying data, conditional formatting based on thresholds, multiple data points woven into a coherent explanation, and data-driven alerts. Power BI's tabular model and Oracle's semantic layer approach data very differently, so there's no one-to-one import path. Replicating narrative views well requires understanding both platforms and deliberately choosing the right Power BI technique, treating narratives as a real component of the migration plan rather than an afterthought bolted on at the end.

A basic strategy before implementation: inventory every existing narrative element and its business purpose, assess each one's complexity and data dependencies, decide which Power BI technique fits each type, and establish a testing method to confirm narrative accuracy once migrated- the same inventory-first discipline covered in our OBIEE to Power BI migration guide for the broader platform migration this work usually sits inside.

Building the Data Model First

Proper data modeling is the foundation everything else depends on. Converting OBI subject areas into Power BI datasets means ensuring the new model represents every data element a narrative actually needs, since Power BI's tabular model works differently from Oracle's semantic layer.

Practical modeling steps that directly support narrative functionality:

Create calculated columns for narrative components used repeatedly across reports.
Design measures that output text, not just numbers, where the narrative calls for it.
Implement row-level security aligned with narrative visibility requirements specifically, not just the surrounding visuals.
Optimize relationship cardinality so narrative context resolves correctly.

Getting this right first makes every subsequent narrative implementation straightforward, not a workaround.

Three Ways to Rebuild a Narrative in Power BI

Once the data model is in place, three techniques cover most narrative requirements, depending on how complex the narrative is.

Method 1: DAX measures that return text. The most broadly useful technique: a measure returns a text string instead of a number, evaluated dynamically against the current filter context. A simple sales performance narrative:

Sales Narrative =
IF(
    [Total Sales] > [Sales Target],
    "Sales exceeded target by " & FORMAT([Total Sales] - [Sales Target], "$#,##0"),
    "Sales missed target by " & FORMAT([Sales Target] - [Total Sales], "$#,##0")
)

Display the result in a card visual or text box for a dynamic narrative that updates as filters change.

Method 2: Power BI's native narrative capabilities. A correction worth being explicit about: older migration guidance often recommends a third-party custom visual called "Narrative Science" for AI-generated narrative text. Microsoft removed that visual from the Power BI Custom Visuals Marketplace years ago, and current community reports confirm it still can't be found or reinstalled. The current supported path is Power BI's native Smart Narrative visual, which auto-generates data-driven narrative summaries built into the platform, alongside Copilot-powered narrative summaries in current Power BI releases. Neither requires a third-party marketplace visual or its licensing.

Method 3: R and Python script visuals. For narratives requiring genuinely advanced text generation, custom R or Python scripts embedded in Power BI visuals offer near-unlimited flexibility, generating narrative text from dataset values and responding to filter context changes. This needs more advanced skills to build and maintain, so reserve it for narratives the other two methods genuinely can't handle.

Security and Governance for Narrative Content

Narrative text often carries sensitive business commentary, and it can leak detail that the surrounding visual correctly protects. Apply row-level security directly to narrative-generating measures, not just the visuals around them; confirm text outputs don't expose data inappropriately even when the numbers are hidden, and extend data loss prevention policies to exported reports containing narrative content. Map OAS security roles to Power BI workspaces and row-level security explicitly, then test the security implementation specifically against narrative content, since a report that correctly restricts a chart can still leak the same information through an unguarded narrative measure- the same governance rigor covered in our key checklist for BI modernization.

Performance, Data Connections, and Complex Calculations

Narrative views can meaningfully affect report performance, especially with complex calculations or large text outputs. Pre-calculate narrative components where possible rather than generating them at query time, use calculation groups to manage multiple narrative scenarios efficiently, and monitor DAX query performance for text-generating measures specifically.

For complex OBIEE-derived calculations, break the logic into smaller, manageable DAX components, use variables for readability and performance, and match Power BI's formatting functions to Oracle's original text presentation where consistency matters to users.

Data connectivity decisions matter too: weigh DirectQuery against Import mode for narrative-heavy reports based on freshness needs, set refresh schedules appropriately for time-sensitive narratives, and consider composite models where a narrative draws from multiple sources, the same connectivity trade-offs covered in our Tableau to Amazon QuickSight guide for a different platform pairing facing a similar SPICE-vs-live-query decision.

Integration, UX, and Modernization Opportunities

User experience. Design consistent narrative presentation patterns across reports, train users to interpret and interact with narrative elements (not just the surrounding visuals), and gather feedback to iterate on implementations.

Visual integration. Narratives work best tightly coupled with the visuals they explain: use bookmarks and buttons to toggle between visual and narrative views, drill-through pages for narrative detail, and tooltip pages to embed narrative context directly within a visualization.

Broader integration. Narrative-containing reports need to work within Microsoft Office, embedded portals, and automated distribution through Power Automate, not just live inside Power BI itself.

Modernization, not just replication. This migration is a real opportunity to exceed what OAS narrative views originally offered: Copilot-driven narrative generation, mobile-optimized narrative layouts, natural-language Q&A for ad-hoc narratives, and cross-report drill-through for deeper context are all genuinely new capabilities OAS didn't have.

Case Study: A Healthcare Narrative Migration

The following is an illustrative example, not an account of a specific customer engagement. No customer names, figures, or performance results are implied.

A healthcare provider migrates over 200 reports containing complex narrative views as part of a broader OAS-to-Power BI migration. The approach centers on building a library of standardized DAX expressions for common clinical and operational narratives, implementing row-level security that specifically preserves patient data privacy in narrative outputs (not just the surrounding charts), using Python script visuals for the small number of genuinely complex clinical narratives that DAX alone can't handle, and establishing a center of excellence to support narrative implementations across the wider report estate going forward.

The likely outcome of an approach like this: narrative functionality preserved without gaps, security validated specifically against narrative content rather than assumed to inherit from the visuals around it, and a reusable DAX library that speeds every subsequent narrative built afterward.

Frequently Asked Questions

Is there a direct Power BI equivalent to OAS narrative views?

No single one. Depending on complexity, you can rebuild narrative views using DAX text measures, Power BI's native Smart Narrative visual, Copilot-generated summaries, or R/Python script visuals for advanced cases.

Is the "Narrative Science" custom visual still available for Power BI?

No. It was removed from the Power BI Custom Visuals Marketplace years ago and remains unavailable, confirmed by recent community reports. Power BI's native Smart Narrative visual and Copilot summaries are the current, supported alternative.

Can DAX measures really return text instead of numbers?

Yes. A DAX measure can return a text string, evaluated dynamically against the report's current filter context, making it one of the most reliable techniques for conditional, data-driven narrative content.

Does row-level security automatically protect narrative text?

Not automatically. Apply row-level security directly to narrative-generating measures and test it against narrative content, since text output can expose data that the surrounding chart correctly hides.

Should DirectQuery or Import mode be used for narrative-heavy reports?

It depends on freshness needs and data volume. DirectQuery suits narratives that need live data; Import mode suits narratives where performance matters more than up-to-the-minute freshness. Composite models are an option when a narrative draws from multiple sources with different needs.

How DataTerrain Accelerates Your Migration

DataTerrain helps organizations navigate OAS-to-Power BI migrations, with specific expertise in replicating and modernizing narrative view functionality. Our approach includes pre-built DAX templates for common narrative scenarios, accelerators for security mapping, and testing frameworks built specifically around narrative accuracy validation, the same broad platform coverage reflected in our OBIEE to Power BI migration guide for the wider migration this work sits inside.

Talk to our migration team