Creating Datasets

After completing this lesson, you will be able to:

After completing this lesson, you will be able to:

  • Explain the types of datasets and the difference between datasets and models
  • Create a dataset


A dataset is a simple collection of data, usually presented in a table. You can use a dataset as the basis for your story, and as a data source for Smart Predict. Datasets are a good choice when you want to create stories quickly and do not want to get into structure definition or when development does not demand IT governance.

For the most part, datasets acquire data via import connections. The only exception is that a dataset can be built based on a live connection to a HANA data repository.

Standalone (Public) and Embedded (Private) Datasets

Access the Datasets menu option from the SAP Analytics Cloud navigation bar. From there, you can create standalone datasets based on a file or an import data source.

Standalone (Public) Datasets

Standalone datasets are data sources that can be used in multiple stories by any users who have access to them. They can be created directly from the vertical toolbar and are also be used for Smart Predict.

Embedded (Private) Datasets

Embedded datasets are created in and exist in only a single story and cannot be shared or used in other stories or with Smart Predict. However, if you need others to be able to use this dataset, you can convert it to a standalone dataset.


A shareable live dataset can be created for SAP HANA. Live datasets can also be created from a story based on SAP Datasphere.

Datasets are intended to supplement models and be used only for ad-hoc, ungoverned data analysis.

Datasets versus Models

CharacteristicsDatasetsLive ModelsImport Models
PurposeAd-hoc/Smart predictPlanning (BPC Embedded only)Planning and Analysis
Scheduled imports availableNoNoYes
Standalone or embeddedBothStandaloneStandalone
Table1 data tableDimensions and data tableDimensions and data table
Calculated measuresNoNoYes
Data Security in SAP Anlaytics CloudNoNoYes

Create a Standalone (Public) Dataset

Business Example

You have a file with HR data and you need to create a public dataset so that it can be used in multiple stories.

Task Flow

In this practice exercise, you will:

  • Create a dataset based on a file
  • Switch date columns to dimension data types
  • Solve a data type issue

Log in to track your progress & complete quizzes