Dataset in Zluri

A dataset in Zluri is a structured collection of data pulled from various modules across your organization's Zluri account. Datasets are the basis for both standard and custom widgets used in dashboards.

A dataset defines what data is available; a widget defines how it's shown.

---

Relationship Between Datasets and Widgets

• A widget is a visual layer (such as a Bar Chart, Table, or Number) built on top of a dataset.

---

Using a Dataset to Create a Widget

Several widgets can use the same dataset. For example:

Example Use Case

Suppose a dataset tracks monthly users of your AI apps. From it, you can build:

• A Line Chart for weekly usage trends
• A Number widget showing total active users
• A Table listing users by department or application

Dataset Structures: Flat vs. Expanded

Every dataset name ends in -Flat or -Expanded, which tells you how its rows are structured.

1. Flat Dataset

Flat datasets contain one-to-one mapped fields. Each row represents a unique entity with no nesting.

Characteristics:

• One row per record
• No repeating or nested fields
• Simple structure for visualization and aggregation

Best For:

• Number widgets (for example, total users or average apps per user)
• Bar Chart widgets grouped by role or department
• Simple tables with summary data

2. Expanded Dataset

Expanded datasets include normalized or relational fields. One entity (like a user) can span multiple rows because of related data.

Characteristics:

• One-to-many relationships across rows
• Supports granular filtering and grouping
• Requires grouping or a distinct count (the COUNT_DISTINCT aggregate under METRICS) to avoid counting a record more than once

Example Table (columns from App-to-User Mapping-Expanded):

Best For:

• Access reviews (users with multiple apps or roles)
• Multi-dimensional filters (user + app + access)
• Breakdown charts (for example, a stacked Bar Chart)

Dataset Selection Guide

---

Prebuilt Datasets in Zluri

Zluri offers the following prebuilt datasets.

---

How to Use a Dataset to Create a Widget

To build a widget from any dataset:

1. Go to Analytics → Widgets.

2. Click Create New Widget.

3. Under Choose a dataset, select a dataset. Each dataset shows a one-line description. You can select only one dataset per widget.

   The Choose a dataset list, with each dataset's description:

   

4. Under Choose chart type, select the chart, then click Create new widget.

5. Configure the widget. See Widgets.

6. Save the widget and add it to a dashboard. See Create a dashboard.

Best Practices

• Use Flat datasets for fast rollups and KPI-level summaries
• Use Expanded datasets for deep-dive visualizations and filtering
• In expanded datasets, always use grouping or the COUNT_DISTINCT aggregate to avoid overcounts. In the widget editor, click METRICS and, in the My metric panel, choose COUNT_DISTINCT under AGGREGATE. For example, to count unique users, use COUNT_DISTINCT on a user column such as User ID.