Skip to content

Your first Data App

3 min read · Updated 2 Nov 2026

This page is the shortest path from nothing to a published Data App. Each step ends with something you can see, and links to the page that covers it in full.

  • You need an Astrato account that can create content.
  • To publish in step 5, publishing must be enabled for your environment and you need Publish permission on the app. Without it you can still build and preview. See Manage Data Apps.
  • You need credentials for a data source, such as Snowflake, BigQuery or Databricks.
  • The Data App Builder is in Private Preview and is off by default. If Data App is missing from the Create menu in step 4, ask your Astrato administrator or Astrato contact. To build with your own coding assistant instead, see Choose how to build.

Open Data & AI → Data Connections, select New connection, choose your source and enter its credentials. Select Validate Connection, then Save connection.

The New connection panel listing the available sources, including Snowflake, Google BigQuery, ClickHouse, Databricks, Amazon Redshift, Supabase, DuckDB, Dremio and PostgreSQL

You are done when the connection appears in the Data Connections list. See Data connections.

Open Data & AI → Semantic Layers, select Create, then New Semantic Layer. Choose your connection as the Datasource and select Create. The empty layer opens in the editor.

The Create menu open, listing New Semantic Layer and Duplicate existing Semantic Layer under Create, and Snowflake Semantic View and Semantic Layer (YAML) under Import

You are done when the editor opens. See Create a semantic layer.

In the editor, add the tables you want, then define the dimensions and measures people will ask about. Publish the layer when you are ready: the Data App preview uses published dimensions and measures only.

The editor’s Map tab with the Tables & Views panel open, listing the schemas in the connected data source

You are done when the layer is published. See Semantic layer editor and Semantic layer publishing.

Both options start from the semantic layer you published in step 3. To compare them, see Choose how to build.

In Analytics, select Create, then Data App. Enter a Data App name, choose your semantic layer and select Create Data App. Describe the app in the Nash panel, for example “Build a dashboard with key metrics and a chart”. The app takes shape in Preview.

The Create menu in Analytics, with Data App, Workbook and Pinboard under Create, and Workbook under Import.

The “What would you like to build?” form, with a Data App name field, a semantic layer selector and a Create Data App button.

The Data App Builder, with an empty Preview on the left and the Nash chat panel on the right, offering suggestions such as building a dashboard with key metrics and a chart.

You are done when the preview shows your data. See Getting started with Data Apps.

Connect your assistant to Astrato, then ask it to build the app from your published semantic layer. Follow Connect an AI assistant to set up the connection.

To publish an app built this way, use the CLI workflow in Data Apps - AI Assistant Quickstart. The Publish button in step 5 applies only to apps built in App Builder.

You are done when the assistant has built the app and it shows your data.

When you are happy with the preview, select Publish. A new app is private until you publish it. Publishing is in preview and needs an enabled environment and publication permissions. See Manage Data Apps.

The published Data App in the builder: the status beside the Version 1 badge reads Published source version 1, with the Preview and Source files tabs and the Update runtime, history and Publish buttons above a live Airbnb Explorer dashboard with Listings and Avg price figures.