Skip to main content

Visualize Data with dentsu.Composable Visualize

Overview

This article explains how to get started with dentsu.Composable Visualize after installation. You'll learn how to access the application, create and manage dashboards, use Visualize Chat to analyze your data, and turn insights into visualizations. It also provides an overview of dashboard management and the session data available for reporting and analysis.

Start Working in Visualize

After launching Visualize, you can begin exploring your data by creating dashboards and interacting with Visualize Chat. The following steps walk through creating your first dashboard, asking questions about your data, and turning results into charts and visualizations.
  1. Navigate to the app. Because authentication is handled through Snowsight single sign-on, you must be signed in to Snowsight in the same browser using a role that has been granted the application's app_public role.
  2. If prompted, select role and warehouse provided by your Snowflake account administrator.
  3. After a period of inactivity, the application may take 1 to 2 minutes to start. During this time, a brief Waking up message may appear. Once the service is running, subsequent access is typically much faster.
  4. The Visualize home page opens when you sign in. Review the guidance on the page to learn the application workflow, or select Create New Dashboard in the purple panel at the bottom of the screen to get started.
  5. The Dashboard section opens with a new dashboard ready for editing.
  6. Use the Visualize chat on the right side by asking a question in the textbox at the bottom of the chat. You can ask questions related to your tag session data or the functionality of the Visualize app. Suggested or example questions are included to get you started.
  7. Once you have asked Visualize Chat a question, you can review the response, SQL used to return the results, and the results themselves. The Results will default to Data, but you can select Chart to customize how the data is pictured.
  8. Clicking Add to Dashboard generates your graph on the left-hand side of the screen.
  9. Hover over a chart to access additional options, including Edit to modify the chart and Export to download or share the chart and its underlying data.

    Note

    It's a good practice to give your dashboard a descriptive name that reflects the charts and data it contains. To rename the dashboard, select the pencil icon next to the dashboard name and replace the default name.

     

Launching the dentsu.Composable Visualize App

If a default role and warehouse are not assigned to your user account, you will be prompted to select them when opening the application. We recommend using a dedicated Snowflake app role and a medium warehouse. Contact your Snowflake administrator if you need assistance choosing the appropriate role or warehouse.

To improve AI response times, you can increase the warehouse size, though doing so may increase Snowflake compute costs.

Note

Larger warehouses may accrue greater Snowflake compute costs. Additionally, the first visit of the day may be slow. The container suspends when idle to save compute and resumes on the next access; the first request after an idle period takes about 1–2 minutes while it starts.

Understanding & utilizing areas within Visualize 

Home tab

The dentsu.Composable Visualize App opens to the Home screen. The default dashboard occupies most of the workspace, with the navigation menu located on the left. Select Home in the navigation menu to return to this screen at any time.

Navigation Menu

The navigation menu is always present on the left side of each screen. Each section of the navigation menu is described in detail below.

Dashboards

From the navigation menu on the left, click Dashboards to get to the Dashboards screen. This section is where users access dashboards.

  • Dashboards – Add New 
    • Clicking Add New creates a new dashboard.
  • Dashboards – My Dashboards 
    • Under My Dashboards, you can access previously created dashboards.

By default, an example dashboard, Example Dashboard, is provided with several example charts that can help you understand how charts may appear when set up.

Dashboards – Team Dashboards (submenu)

Under Team Dashboards, you can access read-only versions of dashboards which a Visualize Administrator has previously shared as Team Dashboards.

Visualize Chat

The Visualize Chat panel allows you to ask questions about your session data in natural language. Use it to analyze visitor behavior, engagement, session activity, browser usage, visitor types, and interactions with your websites or applications. Visualize can also help explain metrics, terminology, and other aspects of the available data.

To get started, enter a question in the chat box and press Enter. To collapse the chat panel, click the right-pointing triangle () in the upper-right corner of the panel. Click it again to reopen the chat.

Chat Questions

To learn what the chat can do, start by asking, "What can I do?" You can use the chat to explore and analyze your website session data by asking questions in natural language. Here are a few examples:
  • Session counts & trends - "How many sessions did we have last month?"
  • Visitor segmentation - "Break down sessions by visitor type"
  • Browser & traffic analysis - "Which browsers drive the most visits?"
  • Identity resolution - "What's our identity match rate this week?"
  • Visit type comparisons - "Compare new vs. returning visitor sessions"
  • Date range analysis - "Show me daily session trends for Q1 2026"
  • Concepts & methodology - "How does identity resolution work?"
  • Terminology - "What does HMID score mean?"

Chat Answers

Chat answers are provided in three sections:

  1. First, you will see the AI interpretation of the question under the heading Response.
  2. Second, the SQL query that the AI generated to get the results is in the expanding item labeled SQL.
  3. Third, the result of the SQL query execution is in the expanding item labeled Results.

    Note

    If the AI does not understand the question, try to word your question with synonyms or contact dentsu.Composable support to request that they update the semantic view to ensure the terminology used is mapped to the correct data columns.

Settings

From the navigation menu, select Settings to view the warehouse and data source currently configured for the Visualize app. This information is displayed as read-only and cannot be modified from this screen.

User Management

From the navigation menu, select User Management to open the User Management screen. This screen is available only to users assigned the administrator role (app_admin). Administrators can use it to manage dashboard sharing by selecting a user's custom dashboard and controlling its visibility.
  1. If the dashboard is private, select Team Share to make it available to all users on the Team Dashboards screen.
  2. If the dashboard is already shared, select Remove Share to remove it from the Team Dashboards screen for all users.

Note

Only Visualize users which have at least one dashboard created will show in the list of users which have dashboards which can be shared or made private.

Administrator procedures

Run these procedures as ACCOUNTADMIN or a role that has been granted the application's administrator role. Replace <code>{APPLICATION_NAME}</code> with the name of your application.

Suspend and resume services

When the application is inactive, you can suspend its compute pool to reduce Snowflake compute consumption. The service automatically resumes when a user accesses the application. If a user encounters a temporary error while the service is starting, wait a few moments and try again.

To manually manage the service, use <code>core.suspend_service</code> and <code>core.resume_service</code>.

Procedure Purpose
{APPLICATION_NAME}.config.generate_setup_sql(...) Generate the account-specific setup script (see Post Installation set up & Verification).
{APPLICATION_NAME}.config.check_setup() Verify the installation; reports the status of each setup step.
{APPLICATION_NAME}.core.suspend_service() Suspend the app's container service to stop compute while the app is idle.
{APPLICATION_NAME}.core.resume_service() Resume (pre-warm) the service ahead of use, avoiding the first-access cold start.