OAC Data Visualization: Projects, Workbooks & Canvases

Explain the core Data Visualization (DV) building blocks in Oracle Analytics Cloud — projects, datasets, and canvases — and how they differ from RPD-based subject areas.

Data Visualization is OAC's self-service front end. Its vocabulary differs from classic OBIEE Answers, which is why it is tested separately.

An RPD subject area is like a professionally organized reference library where a librarian has already catalogued how every book relates to every other book. A DV project is like spreading your own books and a few borrowed ones out on your desk and figuring out on the fly how they connect for the specific question you're trying to answer right now.

Key Concepts

    Project / Workbook   the top-level container
      Canvas             a page of visualisations
      Canvas             another page
      Dataset            the data the project uses

The building flow is different from Answers. In Answers you pick columns on a Criteria tab and then choose views. In DV you drag a field onto a canvas and a visualisation appears, chosen automatically from what you dragged.

    drag Revenue          -> a number tile
    add Region            -> it becomes a bar chart
    add Month             -> it becomes a line chart

You shape the question by adding fields, and the chart follows. That is the self-service model: explore first, decide the presentation afterwards.

Canvases are pages within one project, so an analysis can be laid out as a narrative — summary, then breakdown, then detail — rather than as separate saved objects.

The two data sources, and the distinction that matters.

    a subject area   the governed RPD model. Curated definitions,
                     row-level security applied by the BI Server.
    a dataset        a file, a database connection, a spreadsheet
                     somebody uploaded.

Both are legitimate and they are not equivalent. A project over a subject area inherits the organisation's definitions and security. A project over an uploaded spreadsheet inherits neither.

That is the governance question behind self-service, and the mature answer is to point DV at the governed subject areas by default — so exploration is free and the numbers still agree — and allow uploaded datasets for genuinely ad hoc work.

Data blending joins a dataset to a subject area within a project, which is the common real use: the governed sales model, plus a target spreadsheet the finance team maintains.

What DV does not replace. It is exploration and presentation. A pixel-perfect invoice is still BI Publisher, and a governed enterprise model is still the RPD. DV sits on top of the model rather than instead of it.

What the interviewer is probing.1. "How does data visualisation differ from classic analyses?" Probing: the positioning. Stalls: "It looks nicer." Moves up: it is self-service and exploratory over datasets a user can bring, where classic analyses are governed reporting over the RPD.

2. "What is the governance question it raises?" Probing: the real issue. Stalls: "None." Moves up: who owns the numbers when anyone can upload a spreadsheet — which is why curated datasets and certified content matter.

3. "What is a canvas?" Probing: the vocabulary. Stalls: "A page." Moves up: one page of visualisations within a workbook, with filters that can apply to it or across the workbook.

4. "When would you still build a classic dashboard?" Probing: judgement. Stalls: "Never." Moves up: for governed, repeatable reporting with row-level security and a stable definition that many people depend on.