In this blog post, I want to summarize the new releases from the Google tools that we use daily in datadice. Therefore, I want to give an overview of the new features of BigQuery, Dataform, Data Studio, Google Analytics, and Google Tag Manager. Furthermore, I will focus on the releases that I consider to be the most important ones, and I will also name some other changes that were made.

If you want to take a closer look, here you can find the Release Notes from BigQuery, Dataform, Data Studio, Google Analytics & Google Tag Manager.

BigQuery

Renamings

Google implemented significant naming changes across several data and analytics services. These updates primarily affect the user interface branding, while technical resources such as APIs, client libraries, CLI commands, and Identity and Access Management (IAM) roles retain their original names.

The key rebranding changes include:

• Dataproc to Managed Service for Apache Spark
• BigLake to Google Cloud Lakehouse (BigLake metastore has been renamed to the Lakehouse runtime catalog)
• Dataplex Universal Catalog to Knowledge Catalog
• Looker Studio to Data Studio (The website and endpoint have changed from lookerstudio.google.com to datastudio.google.com)

BigQuery Graph

BigQuery is a Graph database too now! And yes, it is worth its own blogpost!

On a broad scale, you have the following opportunities:

• Create graphs directly from the existing tables
• Use the Graph visual modeler to build your graphs entirely in the UI
• Use GQL (Graph Query Language) to build up your Graph
• Visualize the Graph in a notebook or in BigQuery Studio
• Using Conversational Analytics on your Graphs

With GQL, you can do the following:

Setting up the nodes and edges of existing tables:

CREATE OR REPLACE PROPERTY GRAPH graph_db.FinGraph
  NODE TABLES (
    graph_db.Account,
    graph_db.Person
  )
  EDGE TABLES (
    graph_db.PersonOwnAccount
      SOURCE KEY (id) REFERENCES Person (id)
      DESTINATION KEY (account_id) REFERENCES Account (id)
      LABEL Owns,
    graph_db.AccountTransferAccount
      SOURCE KEY (id) REFERENCES Account (id)
      DESTINATION KEY (to_id) REFERENCES Account (id)
      LABEL Transfers
)

Querying the graphs and visualizing them:

GRAPH graph_db.FinGraph
MATCH
  p = ((person:Person {name: "Dana"})-[own:Owns]->
  (account:Account)-[transfer:Transfers]->(acount2:Account)<-[own2:Owns]-(person2:Person))
RETURN
  TO_JSON(p) AS path;

To create a Graph in the UI, you need to start here

But this topic is too huge to handle in a short section. The rest you need to find out on your own. Maybe we will write a separate blog post about it.

A starting point for further research can be found here.

Data Transfer improvements

The BigQuery Data Transfer Service has been enhanced to support incremental data transfers for a wider range of database connectors. This allows for just transferring new or updated records rather than pulling the entire dataset every time.

Incremental data transfers are now available for the following data source connectors:

• MySQL
• Oracle
• PostgreSQL
• ServiceNow
• Microsoft SQL Server

To configure an incremental transfer (example for MySQL):

• Navigate to Data Transfers in your BigQuery workspace
• Create a new transfer and select MySQL as the source
• During configuration, set up your connection and dataset details
• Choose the Ingestion type “Incremental”
• Decide if you want to use “Append” or “Upsert” as write mode

With the write mode “Append”, the updated row will be added to the table, and the old one remains. “Upsert” updates the original row, and the original values are gone.

More information can be found here (Example MySQL).

New AI Function

Google has introduced AI.KEY_DRIVERS, a new function that performs contribution analysis directly within your SQL queries. This feature allows users to automatically identify data segments that drive statistically significant changes to a summable metric.

Prepare your data so that it includes your dimension columns, one numeric column containing the summable metric, and one boolean column that indicates whether a row belongs to the interest set (True) or the reference set (False).

An example of a query could be:

WITH InputData AS (
  SELECT
    product_gross_revenue_after_discount AS gross_revenue,
    product_name,
    seller_channel,
    billing_country,
    IF(order_date > '2023-07-01', TRUE, FALSE) AS date_after_h2
  FROM `dataset_name.order_item`
  WHERE EXTRACT(YEAR FROM order_date) = 2023)

SELECT * EXCEPT(product_name, seller_channel)
FROM AI.KEY_DRIVERS(
  TABLE InputData,
  metric_col => 'gross_revenue',
  dimension_cols => ['product_name', 'seller_channel'],
  interest_label_col => 'date_after_h2',
  min_apriori_support => 0
);

The function compares an “interest” dataset against a “reference” dataset to find unexpected differences and calculate the relative impact of specific segments (in our example H1 against H2 of 2023).

The function output provides a driver’s array (the dimensions defining the segment) alongside detailed metrics such as metric_interest, metric_reference, difference, and unexpected_difference

The current limitations are the maximum of 12 dimensions, and it only works with summable metrics.

More information can be found here.

Python UDFs

Several enhancements to Python UDFs are available:

• Vectorized UDFs: You can now create vectorized UDFs utilizing the Apache Arrow RecordBatch interface
• Cloud Monitoring Integration: UDFs automatically export operational metrics such as CPU utilization and memory usage directly to Cloud Monitoring
• Concurrency Control: With the new container_request_concurrency option in the CREATE FUNCTION statement, you can set the maximum number of concurrent requests per container instance
• New Quotas: Python UDFs have some limits, including a maximum of 10 GiB for image storage and a mutation rate of 30 operations per minute
• Cost Visibility: Execution costs can be tracked using the external_service_costs column in the INFORMATION_SCHEMA.JOBS view and within the ExternalServiceCosts field in the Job API.

MFA for Google Ads Transfer

Starting 07.05.2026, the BigQuery Data Transfer Service for Google Ads will strictly require Multi-Factor Authentication.

Key details regarding this update:

• Requirement for New Transfers: This policy applies only to new transfer configurations created after 07.05.2026
• Existing Transfers: Current and active transfer configurations are not affected by this change and will continue to run without immediate modification
• Exemption for Service Accounts: Transfers authorized via Service Accounts are exempt from this MFA requirement
• Security Compliance: Ensure that the user account used to set up a new transfer has 2-Step Verification enabled to prevent setup failures

More information can be found here.

Dataform

Developer Connect for Git

The connection process between Dataform and third-party Git repositories (such as GitHub or GitLab) has been significantly streamlined. With the Developer Connect integration, manual secrets management is no longer required, and Dataform can now support repositories located within privately hosted networks.

To implement this connection:

• Navigate to your Dataform project in the Google Cloud Console
• Open the Settings section and click on “Connect with Git”
• Select the option to connect a repository using Developer Connect
• Click on Repository > Link new repository > Create new connection
• Follow the authentication prompts to link your third-party Git provider
• Configure the networking settings if your repository is hosted in a private network
• Verify the connection to enable automated syncing and version control for your SQLX files

With this approach, you need to authorize Dataform once with your GitHub or GitLab account, and afterwards you just need to choose which repository you want to connect.

More information can be found here.

Data Studio

BigQuery data agents

You can publish BigQuery data agents to Looker Studio now. This integration enables the use of Conversational Analytics within your dashboards, allowing stakeholders to interact with data using natural language.

To enable this feature:

• Create a data agent within BigQuery.
• Use the automated publishing option to send the agent to Looker Studio.
• Access the agent within your Looker Studio report to begin using Conversational Analytics.

Furthermore, Conversational Analytics is now available to all users (not just for the Pro version).

More information about the BigQuery data agents creation can be found here, and about the new Conversational Analytics here.

Google Analytics

Task Assistant

The Task Assistant is a new menu to provide tailored recommendations for optimizing Google Analytics property configurations and improving data collection.

The Task Assistant is in the left navigation menu, above the admin area.

The tasks are grouped in the following sections:

• Get started: Basic setups like Data collection, Google signals, audiences, Key events
• Connect your accounts: Google Ads and Search Console link
• Enhance your reporting: Custom insights and audiences
• Optimise your advertising: Google Ads actions
• Add first-party data: user data, campaign data, measurement protocol
• Fix data issues: If Google Analytics recognizes data issues, potential fixes are shown

More information can be found here.

Google Tag Manager

No further release for Google Tag Manager.

Further Links

This post is part of the Google Data Analytics series from datadice and explains to you every month the newest features in BigQuery, Data Studio, Google Analytics, and Google Tag Manager.

Check out our LinkedIn account to get insights into our daily working life and get important updates about BigQuery, Data Studio, and marketing analytics.

We also started with our own YouTube channel. We talk about important DWH, BigQuery, Data Studio, and many more topics. Check out the channel here.

If you want to learn more about how to use Google Data Studio and take it to the next level in combination with BigQuery, check our Udemy course here.

If you are looking for help to set up a modern and cost-efficient data warehouse or analytical dashboards, send us an email to hello@datadice.io, and we will schedule a call.

Originally published at https://datadice.io.

In this blog post, I want to summarize the new releases from the Google tools that we use daily in datadice. Therefore, I want to give an overview of the new features of BigQuery, Dataform, Looker Studio, Google Analytics, and Google Tag Manager. Furthermore, I will focus on the releases that I consider to be the most important ones, and I will also name some other changes that were made.

If you want to take a closer look, here you can find the Release Notes from BigQuery, Dataform, Looker Studio, Google Analytics & Google Tag Manager.

BigQuery

Syntax changes for VECTOR_SEARCH

Google introduced an alternate syntax for the VECTOR_SEARCH function, specifically optimized for single-vector searches. Previously, the function was primarily designed to handle batch searches by passing a query table. The new syntax allows you to pass a single query value directly using the query_value parameter, avoiding the need to construct a single-row table.

This enhancement provides the following benefits:

• Improved query performance compared to using the batch version for a single row.
• Reduced query runtime and cost for common single-query patterns.
• Less slot time usage and lower per-query charges for vector workloads at scale.

The updated syntax for searching a single vector is structured as follows:

VECTOR_SEARCH(
  TABLE base_table,
  'column_to_search',
  query_value => single_query_value,
  top_k => 5
)

More information can be found here.

Conversational Analytics changes

Google expanded the conversational analytics capabilities within BigQuery. To start conversational analytics, you need to click here:

The most important enhancements include:

• Integration of unstructured data: Through new ObjectRef functions, the AI assistant can now directly access files stored in Google Cloud Storage. This allows you to include documents like PDFs and images in the analysis.
• Expanded BigQuery ML capabilities: The agent supports advanced AI functions, such as AI.GENERATE, AI.FORECAST, and AI.DETECT_ANOMALIES.
• Direct interaction with query results: Users can now initiate a conversation directly concerning the output of their SQL queries within the BigQuery Studio editor.

• Cost-efficient queries via partitioning: The conversational agent automatically recognizes partitioned tables and uses relevant columns, such as date ranges, to generate optimized SQL code. This improves query speed and noticeably reduces execution costs.
• Interactive follow-up questions: The agent delivers context-aware next questions that can be executed with a single click in the Google Cloud Console.

More information can be found here.

BigQuery Data Transfer Service remote MCP Servers

Google has introduced the BigQuery Migration Service MCP (Model Context Protocol) server, designed to assist users with various SQL translation tasks directly. This tool simplifies migrating and adapting SQL code for BigQuery.

With the new MCP server, you can perform the following tasks:

• Translate existing SQL queries from other dialects directly into standard Google SQL syntax.
• Generate Data Definition Language (DDL) statements based on your SQL input queries.
• Obtain detailed explanations for the provided SQL translations to understand the syntax changes and underlying logic better.

More information can be found here.

AI Function changes

Google introduced new parameters for some of the AI functions.

The updates include the following enhancements:

• AI.DETECT_ANOMALIES: You can now set a custom context window to determine exactly how many of the most recent data points the model should evaluate (Parameter = context_window).
• AI.FORECAST: This function now allows you to explicitly specify the latest timestamp to use for forecasting (Parameter = forecast_end_timestamp).
• AI.EVALUATE: Provide a custom context window to define the number of recent data points used by the model (Parameter = context_window). Additionally, the function now outputs the mean absolute scaled error for the analyzed time series (output column = mean_absolute_error).

Further changes

BigQuery got a few more updates I want to mention:

• Global default location: Configure a global default location at either the organization or project level, which will be automatically applied when no specific location is provided or cannot be directly inferred from your request.

ALTER ORGANIZATION SET OPTIONS (
  `default_location` = EU
)

;

ALTER PROJECT SET OPTIONS (
  `default_location` = EU
)

• Visual mapping in query performance: A visual mapping of SQL queries within the query execution graph to help users better understand and debug query performance, including a new heatmap functionality that highlights the specific execution steps consuming the most slot-time.

• Automatically deploy open models: BigQuery ML now allows you to automatically deploy open models to Vertex AI endpoints, offering benefits such as automated Vertex AI resource management, the ability to use Compute Engine reservations, and automatic or immediate model undeployment to optimize costs.
• Migration assessment for Snowflake: There is a new member in the migration assessment family: Snowflake. If you want to migrate your Snowflake infrastructure to BigQuery, you get some support here.

Dataform

No further release for Dataform.

Looker Studio

CSV connector changes

In October 2025, a few enhancements for the CSV connector were announced, but just for Pro users.

These changes are now available to all users. More information can be found here.

Google Analytics

No further release for Google Analytics.

Google Tag Manager

No further release for Google Tag Manager.

Further Links

This post is part of the Google Data Analytics series from datadice and explains to you every month the newest features in BigQuery, Data Studio, Google Analytics, and Google Tag Manager.

Check out our LinkedIn account to get insights into our daily working life and get important updates about BigQuery, Looker Studio, and marketing analytics.

We also started with our own YouTube channel. We talk about important DWH, BigQuery, Looker Studio, and many more topics. Check out the channel here.

If you want to learn more about how to use Google Data Studio and take it to the next level in combination with BigQuery, check our Udemy course here.

If you are looking for help to set up a modern and cost-efficient data warehouse or analytical dashboards, send us an email to hello@datadice.io, and we will schedule a call.

Originally published at https://datadice.io.

In this blog post, I want to summarize the new releases from the Google tools that we use daily in datadice. Therefore, I want to give an overview of the new features of BigQuery, Dataform, Looker Studio, Google Analytics, and Google Tag Manager. Furthermore, I will focus on the releases that I consider to be the most important ones, and I will also name some other changes that were made.

If you want to take a closer look, here you can find the Release Notes from BigQuery, Dataform, Looker Studio, Google Analytics & Google Tag Manager.

BigQuery

Parameterized queries

Users can now configure and execute parameterized queries directly within the BigQuery query editor in the Google Cloud console. This update simplifies the process of writing and testing dynamic SQL.

To pass parameters in the query editor:

• Open the BigQuery query editor in your Google Cloud console
• Click on the “More” option in the query editor toolbar and select “Query settings”
• Locate the “Query parameters” section and click “Add parameter” to define the name, data type, and value
• You can use the configured parameter in the SQL code by prefixing the parameter name with an @ symbol (e.g., @parameter_name)

And this is the corresponding query:

SELECT
  name,
  SUM(number) AS amount
FROM `bigquery-public-data.usa_names.usa_1910_2013`
WHERE
  gender = @gender
  AND year = @year
GROUP BY 1
ORDER BY 2 DESC

In these query parameters, you can just add simple datatypes like String, Integer, Float, Date, …

Parameters with the types array, struct, and ranges can be just created and used in the BigQuery CLI.

Example:

bq query \
--use_legacy_sql=false\
--parameter='gender::F'\
--parameter='states:ARRAY<STRING>:["WA", "WI", "WV", "WY"]'\
'SELECT
    name,
    SUM(number) AS count
  FROM `bigquery-public-data.usa_names.usa_1910_2013`
  WHERE
    gender = @gender
    AND state IN UNNEST(@states)
  GROUP BY name
  ORDER BY 2 DESC
  LIMIT 10;'

More information can be found here.

Global queries

Google has introduced “Global queries,” a new feature that lets you query data distributed across multiple geographic locations with a single SQL statement.

BigQuery automatically handles the cross-region orchestration, running optimized sub-queries in remote regions and securely transferring only the necessary filtered data to the primary execution region to deliver a unified result.

The project where the global query is located and needs data from other projects requires the active enable_global_queries_execution option.

ALTER PROJECT `project_id`
SET OPTIONS (
  `region-europe-west3.enable_global_queries_execution` = TRUE
);

The project from which you get data requires the active enable_global_queries_execution option.

ALTER PROJECT `project_id`
SET OPTIONS (
  `region-europe-west3.enable_global_queries_data_access` = TRUE
);

Note: These global queries can lead to higher costs than expected before. The final costs are the sum of the costs from

• every subquery in the remote locations
• the final query
• Copying data between the regions
• Storing the copied data in the primary region for 8 hours

More information can be found here.

Further changes

• Enhanced AI functions: The AI.GENERATE and AI.GENERATE_TABLE functions now allow you to provide descriptions for fields in your custom output schema, while the AI.CLASSIFY function has been upgraded to support multi-category classification for your inputs.
• Undelete datasets: You can now easily recover a deleted dataset to its exact state at the time of deletion, provided the action is performed within your configured time travel window. More information can be found here.
• Custom glossary for conversational agents: To help conversational analytics agents better interpret your specific prompts, you can now create and review custom glossary terms directly in BigQuery, as well as import business glossary terms from the Dataplex Universal Catalog.

Dataform

No further release for Google Tag Manager.

Looker Studio

Conversational Analytics reasoning

Looker Studio has enhanced its Conversational Analytics tool by providing transparency into how the AI interprets user queries. Users can now access a “Show reasoning” feature that provides a plain-text explanation of the logic used to generate a specific data visualization or response. This includes details on how key terms were mapped to specific dimensions and metrics, which filters were applied, and the duration of the processing time.

To view the analytical reasoning:

• Enter your prompt in the Conversational Analytics interface
• Once the response is generated, click the “Show reasoning” button located above the output
• Click “Hide reasoning” to collapse the detailed explanation again

More information can be found here.

Further changes

• Looker Studio Pro for Google Workspace: Subscriptions for Looker Studio Pro can now be directly purchased and centrally managed through your Google Workspace Admin console. More information can be found here.
• Change Pro billing accounts: Administrators can now seamlessly switch the Google Cloud billing account associated with their Looker Studio Pro subscription without any access downside. More information can be found here.
• Export charts as images: Individual charts can now be easily saved as a PNG file or copied directly to your clipboard for quick external sharing. More information can be found here.

• Direct links to report components: You can now generate and share specific URLs that route users directly to an individual component, such as a chart or control element, within a report. More information can be found here.

• Expanded search parameters: Finding reports and data sources on the homepage is now easier, as results can be specifically filtered by Type, Owner, Location, and Date Modified. More information can be found here.

Google Analytics

New Insights on Home page

Google has introduced a new “Generated insights” feature directly on the Google Analytics Home page. This AI-supported tool summarizes the top three data changes observed since your last visit. The insights highlight key configuration updates, anomalies, and seasonality trends to help users identify significant performance shifts immediately.

To use this feature:

• Navigate to the Home page of your Google Analytics 4 property.
• In the upper section of the page, the Generated Insights are shown

So far, I do not see these insights, maybe it still takes a while until they show up in all properties.

The generated insights on the detailed reports are still shown as well.

Google Tag Manager

Improved event collection for Google Ads

There is an automated enhancement for website event data collection. This update allows Google Ads to capture a wider range of event data directly from websites with minimal manual configuration.

This update should lead to more event which are tracked automatically and/or more data sent with the existing events.

Further Links

This post is part of the Google Data Analytics series from datadice and explains to you every month the newest features in BigQuery, Data Studio, Google Analytics, and Google Tag Manager.

Check out our LinkedIn account to get insights into our daily working life and get important updates about BigQuery, Looker Studio, and marketing analytics.

We also started with our own YouTube channel. We talk about important DWH, BigQuery, Looker Studio, and many more topics. Check out the channel here.

If you want to learn more about how to use Google Data Studio and take it to the next level in combination with BigQuery, check our Udemy course here.

If you are looking for help to set up a modern and cost-efficient data warehouse or analytical dashboards, send us an email to hello@datadice.io, and we will schedule a call.

Originally published at https://datadice.io.

In this blog post, I want to summarize the new releases from the Google tools that we use daily in datadice. Therefore, I want to give an overview of the new features of BigQuery, Dataform, Looker Studio, Google Analytics, and Google Tag Manager. Furthermore, I will focus on the releases that I consider to be the most important ones, and I will also name some other changes that were made.

If you want to take a closer look, here you can find the Release Notes from BigQuery, Dataform, Looker Studio, Google Analytics & Google Tag Manager.

BigQuery

New Transfers

The BigQuery Data Transfer Service has added two new connectors, allowing you to ingest e-commerce and marketing data without custom pipelines.

Shopify Data Transfer directly imports your store data into BigQuery.

• Data: Selectable objects like orders, products, customers, and inventory
• Requirements: Shop URL subdomain, Client ID, and Client Secret (via a Shopify app)

More information can be found here.

Mailchimp Data Transfer natively ingests email marketing data for analysis.

• Data: Objects including campaigns, lists (audiences), subscribers, and reports.
• Requirements: Mailchimp API Key.
• Configuration: You can optionally set a start_date to filter historical data

More information can be found here.

Gemini Cloud Assist

There are two useful updates for Gemini Cloud Assist in BigQuery, helping you both analyze past performance and find the right data assets within your projects.

• Job Analysis: You can ask questions to troubleshoot specific jobs, such as Why was my last query slow? Or why did this job fail?. You can also identify which queries were the most resource-intensive over a specific period.
• Resource Discovery: It is now possible to search for resources across your projects using natural language. For example, you can ask for a specific table’s schema or find tables by content, such as Which tables contain demographic information about new users?

More information can be found here.

JDBC driver

A new, open-source Java Database Connectivity (JDBC) driver for BigQuery is now available. This Google-developed driver allows you to connect your Java applications and tools directly to BigQuery using standard database protocols.

This is particularly useful for integrating BigQuery into existing Java-based ecosystems, legacy applications, or third-party tools that rely on the generic JDBC standard rather than native Google Cloud client libraries.

To set it up, you need to

• Include the driver JAR in your project
• The connection is established using a URL pattern starting with jdbc:bigquery://
• Handle authentication, typically by providing a Service Account key or using Application Default Credentials

While it fully supports standard SQL execution, keep in mind that JDBC is generally synchronous. For high-throughput streaming ingestion or complex asynchronous jobs, the native BigQuery API might still be the better choice.

A detailed guide to setting it up can be found here.

AI function changes

You now have more control when using the AI.IF, AI.SCORE, and AI.CLASSIFY functions in BigQuery.

An optional endpoint parameter has been added, allowing you to specify a particular model for the function to utilize.

• Any generally available or preview Gemini model can be used
• If this argument is omitted, the function will choose the most fitting model (cost-to-quality tradeoff)
• The endpoint parameter works for all 3 functions the same

More information can be found here (e.g., for AI.Score).

Dataform

Structure Project with Dataform API

Google has introduced the ability to organize code assets into a hierarchical structure using the Dataform API. This update allows users to manage their development environment more effectively by grouping related resources into folders and repositories.

To utilize this feature via the Dataform API, follow these steps:

1. Access the Dataform API through e.g. the Google Cloud SDK
2. Use the folders resource to create a logical directory structure within your repository
3. Move existing code assets or create new ones within these specific folder paths to maintain a clean workspace
4. Manage permissions and access levels at different levels of the hierarchy as needed for your team

Here is an example of how to create a folder nested within another folder:

curl -H “Authorization: Bearer $(gcloud auth print-access-token)” \
-H “Content-Type: application/json” \
-X POST \
-d ‘{
“displayName”: “folder_name”,
“containingFolder”: “projects/project_id/locations/europe-west3/folders/parent_folder”
}’ \
“https://dataform.googleapis.com/v1beta1/projects/project_id/locations/europe-west3/folders"

More information about the whole permission controlling and more code examples can be found here.

Looker Studio

Filtering across data sources

We already wrote a Blog Post about this trick in 2021! Here you can find it.

It is possible to filter fields from different data sources with the same control. The solution we explained in the blog post is more like a “hack” because it was not natively supported. And now a separate menu can be used for this.

In the menu Resource > “Manage field names and IDs” you can enable these field overrides.

Note: Changing the field ID for a field that is already in use will break the charts. You need to select the new field in the broken charts, then.

And ensure that all fields with the same field id have the same data type.

Hide charts and components for viewers

A new meaningful feature allows you to control the visibility of individual charts, shapes, and other components within your reports. You can now choose to hide specific elements so they remain visible to editors in Edit mode but disappear for users in View mode.

This feature just really works when at least one editor of the dashboard has a Looker Studio Pro account.

How it works:

• Click on the three-dot menu of a chart
• Select the option “Control Visibility”
• Select an existing group membership variable (Pro user can create a new variable)
• Click on Confirm

If your account is part of the group membership variable, you still see the chart in the view mode. If not, you do not see the chart anymore when you are in the view mode.

There are two main use cases for this that I have currently in mind:

• Workflow Improvements: This is excellent for keeping “work in progress” charts in the report without exposing them to stakeholders.
• Internal Documentation: You can use this to add text boxes with instructions, change logs, or data quality notes that are only meant for the dashboard developers and not the end-users.

More information can be found here.

New Histogram charts

You can now visualize the distribution of your data natively using the new Histogram chart type. Unlike standard bar charts that compare specific categories, histograms group continuous data into intervals.

• Statistical Insights: This is essential for understanding the shape, center, and spread of your datasets, helping you spot outliers or skewness immediately.
• Use Case: A classic example is analyzing Session Duration or Order Value. Instead of just seeing an average, you can see exactly how many users fall into specific time or value ranges (e.g., 0–10s, 11–20s, etc.).

The configuration part is a bit confusing.

When you select the histogram, the “auto bin” option is active, and the categories are chosen from Looker Studio based on the distribution of the metric.

And then you can set up the number of bars yourself in the Bin section.

But the distribution of the bins does not work really well for a lot of fields. Then you need to set up the bins yourself by changing the calculation of the dimension.

More information can be found here.

Google Analytics

Cross-channel budgeting

Google Analytics 4 is taking a big step towards becoming a comprehensive planning platform with the introduction of Cross-Channel Budgeting. This feature moves beyond retrospective reporting, allowing you to forecast performance and optimize media spend directly within the interface.

You can access this new toolset by navigating to Advertising > Planning > Budgeting.

The feature offers two main planning types

• Projection plans: help you monitor if your current campaigns are on track to meet your KPIs.
• Scenario plans: allow for “what-if” analysis, letting you simulate how different budget allocations would impact your future revenue or conversions.

There are some requirements before you can use the Cross-channel budgeting:

• You need at least one year of historical data for both conversions and campaigns.
• Data Requirements: Crucially, the model requires data from at least two channels, specifically covering both Google and non-Google sources. This means you will likely need to use Data Import to bring in cost and campaign data for your non-Google channels.
• Compatibility: The feature relies on your Primary Channel Grouping, and all channels must have associated cost data to be included in the plans.

Especially due to the requirement of “One year of cost data for non-Google integrations” I can not really try out this feature now.

To get more information on how it works and what can be done, you can take a look here.

New report for conversion attribution

Google has introduced the Conversion attribution analysis report within the advertising workspace. This new tool is designed to uncover the full value of marketing channels throughout the entire customer journey.

The report offers two specialized views to provide deeper insights:

• Last Click: This view identifies channels that engaged customers early in their journey but were not the final click. It is particularly useful for justifying spend on upper-funnel channels like YouTube.
• Data-driven attribution: This view categorizes touchpoints into Early, Mid, and Late stages. It specifically separates single-touchpoint paths from multi-touchpoint journeys to distinguish between standalone drivers and campaigns that help close complex journeys.

To access the new report, follow these steps:

1. Navigate to the “Advertising” workspace in your Google Analytics 4 property.
2. Select “Attribution” and then click on “Conversion attribution analysis”.
3. Filter by specific conversion events or date ranges to analyze the performance of your upper and mid-funnel campaigns.
4. Review the “Early”, “Mid”, and “Late” stage data to adjust your budget allocation towards high-impact touchpoints.

Some further options can be set up:

More information can be found here.

Google Tag Manager

Google tag gateway changes

There are 2 new small changes regarding setting up the Google Tag Gateway in GTM. You can configure the Google Tag Gateway in the Admin section of the Server container:

• Akamai can be selected as a CDN, next to Cloudflare

• Using GCP’s Global external Application Load Balancer to set up the Google Tag settings

Further Links

This post is part of the Google Data Analytics series from datadice and explains to you every month the newest features in BigQuery, Data Studio, Google Analytics, and Google Tag Manager.

Check out our LinkedIn account to get insights into our daily working life and get important updates about BigQuery, Looker Studio, and marketing analytics.

We also started with our own YouTube channel. We talk about important DWH, BigQuery, Looker Studio, and many more topics. Check out the channel here.

If you want to learn more about how to use Google Data Studio and take it to the next level in combination with BigQuery, check our Udemy course here.

If you are looking for help to set up a modern and cost-efficient data warehouse or analytical dashboards, send us an email to hello@datadice.io, and we will schedule a call.

Originally published at https://datadice.io.

In this blog post, I want to summarize the new releases from the Google tools that we use daily in datadice. Therefore, I want to give an overview of the new features of BigQuery, Dataform, Looker Studio, Google Analytics, and Google Tag Manager. Furthermore, I will focus on the releases that I consider to be the most important ones, and I will also name some other changes that were made.

If you want to take a closer look, here you can find the Release Notes from BigQuery, Dataform, Looker Studio, Google Analytics & Google Tag Manager.

BigQuery

Autonomous embeddings in Tables

These embeddings are located in an additional column. This column is based on another source column. If the source column changes, based on a model from Vertex AI, the embedded column changes accordingly.

You can create such a table with an embedding by doing the following:

CREATE TABLE embeddings_test.movies (
  title STRING,
  description STRING,
  genre STRING,
  description_embedding STRUCT<result ARRAY<FLOAT64>, status STRING>
    GENERATED ALWAYS AS (AI.EMBED(
      description,
      connection_id => 'eu.vertex_ai_eu_connection',
      endpoint => 'desc-embedding-005'
    ))
    STORED OPTIONS( asynchronous = TRUE )
);

Then fill the table with some data. Just the 3 columns title, description and genre need to be filled, the description_embedding gets created based on these.

Additionally, a vector index can be used to speed up the system, but we will skip this part.

Then you can ask the AI questions about your data, and you get a distance for every table entry. The smaller the distance, the better the match.

SELECT base.title, base.description, distance
FROM AI.SEARCH(TABLE embeddings_test.movies, 'description', "The movie contains horror parts")
ORDER BY 3

Further information can be found here.

New Transfers

Several new Data Transfers are now available to ingest data into your BigQuery project, including:

• PostgreSQL
• Microsoft SQL Server
• MySQL
• Oracle
• Klaviyo
• HubSpot

The configuration for the relational databases (PostgreSQL, Microsoft SQL Server, MySQL, and Oracle) works similarly. You will need to provide the host, port number, database name, username, and password. There are also multiple TLS options available for securing the connection.

Once the connection is established, you can browse the database to select the required tables or add them manually.

Once the connection is established, you can browse the database to select the required tables or add them manually.

For the SaaS Connections:

• HubSpot: Requires an Access Token
• Klaviyo: Requires a Private API Key

For both services, you can select specific objects to transfer.

Additionally, it is now possible to ingest data from blob storage (Amazon S3, Azure Blob Storage, Google Cloud Storage) into Iceberg tables in BigQuery.

Dataform

Private Workspaces

There is a new kind of Dataform workspaces available: Private Workspaces. If this option is active, only the creator of the Workspace sees the Workspace. Just this account has access to the code changes, compilation errors, and logs for the compilations and invocations from the workspace.

So far, there is no button to activate this option, probably because the feature is in preview.

Instructions on how to set it up already, can be found here.

actAs Permission

Strict act-as mode enforcement is approaching. Please be prepared: The Service Account executing the scripts requires the iam.serviceAccounts.actAs permission. Google will enforce this requirement in the coming weeks.

You can monitor Cloud Logging, which now provides improved details if errors occur due to the Strict act-as mode.

Instructions on how to verify your Service Account setup can be found here.

Looker Studio

Changes for Scheduled Reports

For a long time, you have been able to send your dashboard pages to certain accounts, for example via email.

So far, the shortest period of sending the dashboard data was daily. Now it is even possible to send the dashboard hourly. This feature is just available for Pro users.

If you update your data on an hourly basis, this new feature will be very useful — for example, if you want to track performance during an active sale.

A second update can be found in the schedule under “Advanced options.” The new option, “Expand tables to show all rows (PDF),” ensures that all table rows are displayed, up to a limit of 2,000 rows.

Please use this option with caution: depending on your active filters, it may result in a PDF with a very large number of pages.

Google Analytics

Analytics Advisor

This new Analytics Advisor is at the end … Gemini for your Google Analytics property.

So you can ask any question there regarding your Google Analytics data, and you get the corresponding answer. It can create tables and charts as well, and can link to further reports if it fits.

This kind of question-answer game was already available in some parts before in Google Analytics, but now it is always available and unified in the Analytics Advisor.

You can access this advisor on any page in the top right corner.

New audience templates

There are two new audience templates available:

• High-Value Purchasers: Includes the top X percentile of LTV users
• Disengaged Purchasers: Contains users who have not purchased in the last X days

The new audiences can be used in Google Analytics and be exported to Google Ads as well, to use especially in customer lifecycle goals.

Google Tag Manager

New built-in variables

There are 3 new built-in variables available in GTM:

• Analytics Client ID
• Analytics Session ID
• Analytics Session Number

They contain the current format of these 3 values, which Google Analytics uses to identify the current user and session. If multiple Measurement IDs are active on the website, all different values per ID are saved.

Rebuilding this was always highly unstable and connected with a lot of effort.

One Use Case would be:

If you track data using the Measurement Protocol, we were never able to fully replicate the GA4 session logic. These new variables might solve the issue, but we haven’t tested them yet.

Furthermore, you can push this data to another system to compare the data with GA4.

And there is one new Analytics Storage variable type available. Here you can define which value you want and from which Measurement ID to get the single value.

Further Links

This post is part of the Google Data Analytics series from datadice and explains to you every month the newest features in BigQuery, Data Studio, Google Analytics, and Google Tag Manager.

Check out our LinkedIn account to get insights into our daily working life and get important updates about BigQuery, Looker Studio, and marketing analytics.

We also started with our own YouTube channel. We talk about important DWH, BigQuery, Looker Studio, and many more topics. Check out the channel here.

If you want to learn more about how to use Google Data Studio and take it to the next level in combination with BigQuery, check our Udemy course here.

If you are looking for help to set up a modern and cost-efficient data warehouse or analytical dashboards, send us an email to hello@datadice.io, and we will schedule a call.

In this blog post, I want to summarize the new releases from the Google tools that we use daily in datadice. Therefore, I want to give an overview of the new features of BigQuery, Dataform, Looker Studio, Google Analytics, and Google Tag Manager. Furthermore, I will focus on the releases that I consider to be the most important ones, and I will also name some other changes that were made.

If you want to take a closer look, here you can find the Release Notes from BigQuery, Dataform, Looker Studio, Google Analytics & Google Tag Manager.

BigQuery

Error handling support with Gemini

Gemini now offers even better support for your SQL coding. You can use it to fix errors either before or after running your query.

Before execution: Highlight the problematic code and select Refine > Fix it. A chat window will appear with a suggested solution. Simply click Apply to update the SQL code.

After execution: If BigQuery shows an error, click the “Gemini suggested fixes” button. The chat will open and explain the fix. Clicking Apply & Run automatically updates the code and re-executes the query.

JSON_FLATTEN function

This new function does exactly what its name implies: it flattens the JSON data passed as the first parameter.

The function recursively traverses a JSON object (even through arrays) and extracts all leaf nodes. It effectively transforms a hierarchical tree structure into a flat list of values.

SELECT JSON_FLATTEN(JSON '[[[[1], 2], 3], {"earth": 3}, false]') AS json_flatten

Use Gemini 3.0 in generative AI functions

BigQuery got a lot of easy-to-use AI functions for your SQL Code. Some of them I explained in our last blog post.

Furthermore, you probably heard about the new Gemini 3.0. The new, most powerful version of Gemini.

Gemini 3.0 is not the default version for the AI functions yet. To use it, you need to define the endpoint in the function. The corresponding endpoint URL is (The project ID needs to be replaced):

https://aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/global/publishers/google/models/gemini-3-pro-preview

Then you can do the following:

SELECT
  title,
  movie_url,
  review,
  AI.CLASSIFY(
    review,
    categories => ['negative', 'neutral', 'positive'],
    connection_id => 'us.vertex_ai_us_connection',
    endpoint => 'https://aiplatform.googleapis.com/v1/projects/datadice/locations/global/publishers/google/models/gemini-3-pro-preview') AS category
FROM `bigquery-public-data.imdb.reviews`

BigQuery ML changes

BigQuery ML supports the TimesFM 2.5 time series foundational model now.

Additionally, it supports a lot of AI functions now:

• AI.DETECT_ANOMALIES
• AI.GENERATE_EMBEDDING
• AI.GENERATE_TEXT
• AI.GENERATE_TABLE
• AI.GENERATE_BOOL
• AI.GENERATE_DOUBLE
• AI.GENERATE_INT
• AI.SIMILARITY
• AI.EMBED
• AI.GENERATE

Other changes

• Folders can be created in BigQuery to organize saved queries, canvases, notebooks, and more

• The query execution graph is able to show a query text heatmap. Further information can be found here
• The following SQL dialects can be translated to GoogleSQL: Apache Impala SQL and GoogleSQL. Translation of GoogleSQL can be used for improvements of the translated code from another dialect
• A new BigQuery Agent Analytics plugin in the ADK is available to export the data from the agent interactions to BigQuery. Further information can be found here

Dataform

Small Changes

Google just released minor improvements to Dataform:

• Support for custom constraints to allow better control over resources like “CompilationResult” or “WorkflowInvocation”
• Creation of BigLake tables for Apache Iceberg in BigQuery

Looker Studio

Conditional Formatting on bar charts

Conditional formatting is now available for bar charts. You can choose between two types:

• Rules: Change colors based on specific criteria
• Gradients: Adjust color intensity based on the value

With the Rules type, you can customize the appearance of bars and data labels. If you apply the rule to the entire row and the condition is met, all bars associated with that dimension are colored accordingly.

The remaining settings are identical to those for tables.

Cloud Audit Logs (Pro)

Information about administrative activities and data access is logged automatically and can be viewed in the Logs Explorer of the corresponding Google Cloud Project now. Logged events include:

• Permission Management
• Ownership Transfers of dashboards or data sources
• API calls

In the Logs Explorer, you can find the logs under “Audited resource”:

Parameters as default values in filters

Parameters are a good opportunity to build a more complex dashboard with the help of the user’s input.

You can use the parameter now to define the standard value of a filter. The value of the parameter is used then, when the user did not choose a value for the filter.

As shown in the screenshot below, the parameter can have a default value as well, which is used in the filter too.

Google Analytics

User-provided data improvements

You can send additional user data (like email or name) to Google Analytics to enable more reliable user analysis.

Google has improved the analysis capabilities of this feature to deliver better performance, specifically:

• A focus on activation and Ads Conversions
• Improvements to Enhanced Conversions and Customer Match Audiences
• More accurate attribution and event reporting

Additionally, Google launched a new UPD infrastructure. This is currently used for customers who start sending UPD data. Customers with existing UPD setups will be migrated to the new infrastructure in Q2 2026.

Google Tag Manager

No further release for Google Tag Manager.

Further Links

This post is part of the Google Data Analytics series from datadice and explains to you every month the newest features in BigQuery, Data Studio, Google Analytics, and Google Tag Manager.

Check out our LinkedIn account to get insights into our daily working life and get important updates about BigQuery, Looker Studio, and marketing analytics.

We also started with our own YouTube channel. We talk about important DWH, BigQuery, Looker Studio, and many more topics. Check out the channel here.

If you want to learn more about how to use Google Data Studio and take it to the next level in combination with BigQuery, check our Udemy course here.

If you are looking for help to set up a modern and cost-efficient data warehouse or analytical dashboards, send us an email to hello@datadice.io, and we will schedule a call.

In this blog post, I want to summarize the new releases from the Google tools that we use daily in datadice. Therefore, I want to give an overview of the new features of BigQuery, Dataform, Looker Studio, Google Analytics, and Google Tag Manager. Furthermore, I will focus on the releases that I consider to be the most important ones, and I will also name some other changes that were made.

If you want to take a closer look, here you can find the Release Notes from BigQuery, Dataform, Looker Studio, Google Analytics & Google Tag Manager.

BigQuery

New Data Transfers

There are new data sources available, where you can get the data from:

• PayPal
• Stripe

From PayPal, you can access Transactions, Disputes, Payments, Balance, Products, and Invoices. There are some things to know and some limitations, which can be found here.

From Stripe, you can get a lot of data as well. The guide can be found here.

Additionally, data from Google Analytics 4 can be imported as well. This Transfer imports aggregated data. Per Transfer, you can import one defined report. You can select a maximum of 9 dimensions and 10 metrics per Transfer.

Note: Some dimensions and metrics are not compatible with each other, as usual for GA4. With the support of the GA4 Dimensions & Metrics Explorer, you can check if your selected combination is valid.

Embed natural language as comments

Access to AI features gets easier every day. Using Gemini to create your BigQuery Query is just one comment away.

You can do the following now:

• Open a new SQL Query editor
• Write in a comment the analysis you want to do
• Click enter
• Wait until a SQL query is suggested
• Click Tab to accept the suggested query

Navigation changes in BigQuery Studio

BigQuery Studio got a new navigation menu. The explorer pane is now separated into three sections: Explorer, Classic Explorer, and Git repository.

The Classic Explorer is still the default choice, but the other sections can be selected at the top level.

You can still search in the new Explorer, but the search results are shown in a separate tab. Additionally, the Job history is part of the menu now, which opens a new tab with a list of the most recent jobs.

Furthermore, the way a new tab opens is completely new.

• Clicking on a resource opens it on the same tab
• By clicking on a resource with Ctrl / Cmd it opens in a new tab
• By default, the tab name is italicized. If you double-click on the name, it changes to a normal font, and then this tab will not be replaced

And the third menu from the screenshot above is a dedicated tab for the linked repositories with BigQuery.

Jobs detail page

In the Jobs Explorer menu, you see a list of the historical BigQuery Jobs.

With the new feature you can click on a BigQuery Job to go to a Job Detail Page and observe the Performance of the Job. But even more interesting, you can compare two jobs with each other:

Then, select the jobs you want to compare. Unfortunately, you cannot search or filter the jobs list yet.

After the selection, you see the key metrics of both jobs next to each other. Apart from that, you can look at the SQL Queries and Execution graphs too.

Data Engineering Agent

The new Data Engineering Agent in BigQuery can be used to build pipelines with AI support.

When you create a new pipeline, you can click on “Ask agent” and chat with the agent. The agent merely proposes a pipeline configuration. You must apply these changes to activate them.

As you can see in the “Ask agent” modal, you can add instruction files. This helps the agent in creating the pipeline correctly.

Managed AI functions

These are really nice new features!

There are 3 easy-to-use AI functions you can use in every SQL Query in BigQuery.

For all these AI functions, you need to set up a BigQuery to Vertex AI connection. A guide can be found here.

The first one is AI.CLASSIFY:

SELECT
  title,
  movie_url,
  review,
  AI.CLASSIFY(
    review,
    categories => ['negative', 'neutral', 'positive'],
    connection_id => 'us.vertex_ai_us_connection') AS category
FROM `bigquery-public-data.imdb.reviews`

• First input parameter: string input value (review)
• Second input parameter: a certain number of categories the AI can choose from
• Third input parameter: the connection_id
• The fourth column of the output table represents the sentiment of the review: “negative, neutral, or positive”

The second function is AI.SCORE:

SELECT
  AI.SCORE((
    """
    On a scale from 1 to 10, rate how much the reviewer liked the movie.
    Review:
    """, review),
    connection_id => 'us.vertex_ai_us_connection') AS ai_rating,
  reviewer_rating AS human_rating,
  review
FROM `bigquery-public-data.imdb.reviews`

• First input parameter: the prompt for the AI
• Second input parameter: the connection_id
• The first column of the output table contains the AI-predicted rating, while the second column shows the actual user rating

The third function is AI.IF:

SELECT
  title, 
  review,
  AI.IF(
    ('The review contains spoiler of the movie: ', review),
    connection_id => 'us.vertex_ai_us_connection') AS review_contains_spoiler
FROM `bigquery-public-data.imdb.reviews`

• First input parameter: the prompt for the AI
• Second input parameter: the connection_id
• The table’s third column contains a BOOLEAN value indicating whether the AI detected spoilers in the review

BigQuery ML changes

Again I want to mention the changes happened to BigQuery ML:

• Contains a TimesFM univariate time series forecasting model now and can be used with the two functions AI.FORECAST and AI.EVALUATE
• Manage open models as Vertex AI endpoints, which has some benefits you can find here

Other changes

There are a lot of other BigQuery changes I want to mention:

• Creation of reservation groups is possible, to have better control of slot allocation inside the group
• Row capacity for Pivot tables in Connected Google Sheets is increased from 100.000 to 200.000
• There is a request form available to sign in for early access to Conversational Analytics in BigQuery. When you get access, you can “chat” with your data and create agents

Dataform

Set the priority of BigQuery jobs

In Dataform, there is a new option available for Workflow configurations and manual executions. You can run these jobs as high-priority jobs.

This option is activated by default for all existing workflows. High-priority jobs run immediately when they are triggered, independent of how big the workload is (from other jobs) at this time.

If the option is not activated and other jobs are running in parallel, this job will wait until the other jobs are completed.

Looker Studio

Schedule reports with Slack (Pro)

You can already send your dashboards around via e-mail, Google Chat, and more. This family got a new member: Slack!

The settings are the same as for Email Scheduling. You can select which pages you want to send and the delivery frequency. Specify the destination for the report by choosing a Slack channel or user.

We currently have some problems linking Slack with Looker Studio. Therefore, here is a guide on how to set it up.

Google Analytics

Import Meta and TikTok Costs

Google adds more and more importers where you can get the cost data from. The goal is for Google Analytics to evolve into a central analysis hub where all key data is consolidated.

You can now add costs from Meta and TikTok, too. When the import starts, the last 24 months of historical data are imported automatically. If you need more data from before, you have to import the cost data manually.

This kind of import is just recommended when you use the same utm_source and utm_medium across your campaigns. If not, reporting problems can occur, and importing the cost data manually is the recommended way then.

Benchmarks for more metrics

In your Google Analytics property, you get on the overview page some benchmark values for a certain number of metrics, like “New user rate”.

Before the change, it was limited to percentage or ratio metrics. Now the benchmarks for 20 metrics were added, which are absolute numbers like “New Users”.

Google Tag Manager

No further release for Google Tag Manager.

Further Links

This post is part of the Google Data Analytics series from datadice and explains to you every month the newest features in BigQuery, Data Studio, Google Analytics, and Google Tag Manager.

Check out our LinkedIn account to get insights into our daily working life and get important updates about BigQuery, Looker Studio, and marketing analytics.

We also started with our own YouTube channel. We talk about important DWH, BigQuery, Looker Studio, and many more topics. Check out the channel here.

If you want to learn more about how to use Google Data Studio and take it to the next level in combination with BigQuery, check our Udemy course here.

If you are looking for help to set up a modern and cost-efficient data warehouse or analytical dashboards, send us an email to hello@datadice.io, and we will schedule a call.

In this blog post, I want to summarize the new releases from the Google tools that we use daily in datadice. Therefore, I want to give an overview of the new features of BigQuery, Dataform, Looker Studio, Google Analytics, and Google Tag Manager. Furthermore, I will focus on the releases that I consider to be the most important ones, and I will also name some other changes that were made.

If you want to take a closer look, here you can find the Release Notes from BigQuery, Dataform, Looker Studio, Google Analytics & Google Tag Manager.

BigQuery

Data preparation changes

Google is placing a strong emphasis on BigQuery’s data preparation capabilities, aiming to decrease the reliance on SQL models and make data analysis more accessible to a wider audience.

Now you can do the following with data preparation:

• Loading files from Cloud Storage into data preparation

• Unnesting arrays (Putting each array element into a single row)
• Flattening records

Data Science Agent changes

The Data Science Agent is the other new and prominent tool in BigQuery that has a lot of new features. The Data Science Agent itself I explained in my former blog post.

The Agent got some improvements in the last days:

• Using BigQuery ML and DataFrames in your prompts
• Adding a BigQuery table with an @ or a file with a + to your prompt

• Using Apache Spark or PySpark keywords in your prompt

Destination table node in Data canvas

The data canvas in BigQuery can be used to build chains of data models. You can include joins, custom queries, and many more.

Now there is the possibility to add a data destination node, which you can use after a query node. After setting up the destination dataset and table, you can select “Keep in Sync”. When it is active, the former query node overrides the destination table automatically.

Options like Table Partitioning or Clustering can not be set up.

Gemini CLI extension

For a long time, a Gemini CLI has been available for your device to use the AI capabilities. The CLI got an extension to connect to your GCP Project and ask Gemini about your tables and data.

There is a Data Analytics Extension and a Conversational Analytics Extension available:

Data Analytics Extension:

• Finding Tables in BigQuery
• Execute SQL models
• Do forecasting or contribution analysis

Conversational Analytics Extension:

• Asking complex questions about the data
• A server-side analytics agent comes into action here

On the cost side, you need to pay for the tables Gemini reads (based on the generated SQL queries), but there are no additional costs (for now).

A guide on how to use the Gemini CLI can be found here.

Add Tables and Views to BigQuery pipelines

We already made a detailed blog post about BigQuery Pipelines (formerly called Workflows) here.

Pipelines were able to connect new or existing Queries, Notebooks and Data preparation. New members of this family are tables and views. You can build the following:

When you set up a view or table, you can just enter the name, but not the dataset. The default dataset is “dataform”. You can change it afterwards in the details of the table/view.

The desired output of the table/view you need to define in the corresponding tasks, so it works differently as mentioned earlier for the data canvas update, where you just define the output table.

Changes to the table/view models are always automatically saved.

BigQuery ML changes

There are some new BigQuery ML changes available, which I want to mention:

• Visualisation of model monitoring is available. More information can be found here.
• Different Embedded Models are available and can be used with the function “ML.GENERATE_EMBEDDING”
• Performing supervised tuning on a remote model with gemini-2.5-pro or gemini-2.5-flash-lite model

Dataform

Update Incremental Tables

Just a small change:

If you have an incremental table schema for one model, you can update it without a full table refresh.

Automatic Processing Location

In the repository settings in the workflow_settings.yaml file, you can set a default execution location:

defaultLocation: europe-west3

It is not mandatory anymore, because if there is no defaultLocation, it takes the location from the datasets that are referenced in the SQL code.

Looker Studio

Treemap charts improvements

There are some new features for the Treemap chart:

• There is a separate field to choose which dimension or metric defines the colour

• Changing the colour shades
• Change the border radius for the shapes (shown in the upper screenshot)
• Choosing different Label options

Sorting charts

Tables in Looker Studio dashboards can now be sorted by up to 10 fields.

All fields of the data source can be used, not just the ones which are used in the table.

Google Analytics

Import Pinterest and Snapchat Costs

There are new cost data import options available directly in GA4.

You can add Pinterest and Snap Ads costs now. One limitation is that only the data of the last 24 months gets imported.

If you have already imported data with the manual approach, you need to take care that you have no duplicate cost data in the property. No automatic deduplication is happening then.

A guide to importing Pinterest data can be found here, and for Snapchat here.

Google Tag Manager

No further release for Google Tag Manager.

Further Links

This post is part of the Google Data Analytics series from datadice and explains to you every month the newest features in BigQuery, Data Studio, Google Analytics, and Google Tag Manager.

Check out our LinkedIn account to get insights into our daily working life and get important updates about BigQuery, Looker Studio, and marketing analytics.

We also started with our own YouTube channel. We talk about important DWH, BigQuery, Looker Studio, and many more topics. Check out the channel here.

If you want to learn more about how to use Google Data Studio and take it to the next level in combination with BigQuery, check our Udemy course here.

If you are looking for help to set up a modern and cost-efficient data warehouse or analytical dashboards, send us an email to hello@datadice.io, and we will schedule a call.

Latest Updates on Google Data Analytics (September 2025) was originally published in Geek Culture on Medium, where people are continuing the conversation by highlighting and responding to this story.

In this blog post, I want to summarize the new releases from the Google tools that we use daily in datadice. Therefore, I want to give an overview of the new features of BigQuery, Dataform, Looker Studio, Google Analytics, and Google Tag Manager. Furthermore, I will focus on the releases that I consider to be the most important ones, and I will also name some other changes that were made.

If you want to take a closer look, here you can find the Release Notes from BigQuery, Dataform, Looker Studio, Google Analytics & Google Tag Manager.

BigQuery

Data Science Agent

Google is adding more and more ways to interact with your data by using AI. With the new Data Science Agent, you can write certain Gemini prompts against your data, and the AI will create the corresponding analysis in a Notebook.

How to start the Agent:

Gemini creates a list of todos it would handle, and if you accept them, it will start with the analysis:

Every time it creates a request for the data, you need to accept it:

And at the end, it creates a summary with the insights:

Chained function call

A feature to make your queries more readable, but it is just an optional way of coding.

You can chain multiple functions together, where the order is important and the single parts are separated by a dot.

Example:

SELECT
  (' Mercury Saturn Earth Mars ')
  .REPLACE('Saturn', 'Venus')
  .CONCAT('Jupiter')
  .TRIM()
  .CONCAT (' Saturn')

-> "Mercury Venus Earth Mars Jupiter Saturn"

There are some limitations or requirements for chained functions you can find here.

Temporary variables

With this new feature, you can set up variables just for one column calculation. It is called WITH expressions. The WITH clause to create temporary tables is a different thing and should not be mixed up.

The created variable(s) in the WITH expression are just available in this expression/column.

Example:

SELECT WITH(x AS 'mercury ',
            y AS CONCAT(x, 'earth '),
            z AS REPLACE(y, 'earth ', 'venus '),
            CONCAT(x, z)) AS result,

-> "mercury mercury venus"

You can define a variable to use it multiple times in the calculation of the value.

The following example leads to an error, because the variable is not available for the column “result_2”:

SELECT WITH(x AS 'mercury ',
            y AS CONCAT(x, 'earth '),
            z AS REPLACE(y, 'earth ', 'venus '),
            CONCAT(x, z)) AS result,
            y AS result_2

-> Unrecognized name: y

Show data in a map

BigQuery already has some visualization possibilities to get more insights into your query results. Next to Line, Bar, and Scatter, there is a map visualization available now.

Geographical data can be shown on a map now. Additionally, some basic setups for the map can be made:

• Color Schema
• Opacity
• Choose a data column
• Point size

Example:

New Reference panel

It’s a new small panel on the right side of the SQL Editor. The panel shows up when you click on “Query” on a table.

With the Reference menu, you can insert a new query snippet. It adds the SQL Query “SELECT FROM `project_name.dataset_name.table_name` LIMIT 1000”. By clicking on the columns, you add the fields to the SQL Editor.

Furthermore, you can click on Preview, and a new overlay appears where you can scroll through the data.

Small Changes

BigQuery got a lot of new features in August, so I want to mention some further updates:

• The advanced runtime feature includes the short query optimization feature now
• Query results can be saved to Google Cloud Storage as JSON, Avro, or Parquet

• The data preparation with Gemini support got the “data deduplication” and “data aggregation” features

Dataform

No further release for Dataform.

Looker Studio

Colouring of embedded query results in text

You can enhance your insights in the dashboards by displaying specific query results within text. With the update, conditional formatting for these query results is active now, so they can have a coloured background or font.

The conditional formatting menu works the same as the one for tables.

You need to add the text variable by clicking on it in the text.

Then you can click on “Conditional Formatting”

Data labels for waterfall charts

In waterfall charts, you can customize the labels a bit more now.

You can set up the label position, inside and outside (which did not change anything in my example, probably it depends on different bar relations), and some contrast settings.

Setting up a data label background color

A lot of small new features for Looker Studio.

For data labels in your charts, you can add a background colour to them, so that the data labels are more visible.

In the style tab, you can add a data label to every shown metric.

Responsive reports: Vertical stacking

Multiple parts can be stacked on each other now in responsive reports.

With this section, I want to remind you that the responsive version of Looker Studio is available. These reports reorganize the arrangement of the dashboard elements based on the screen size. In our Blog Post with the March updates, we describe the feature in more detail.

Google Analytics

Ecommerce data availability

Just a small update. Some Item-scoped Ecommerce dimensions are available in more reports now.

• In Reporting sections as a secondary dimension, in filters, comparisons, and customizations
• Ecommerce metrics are available in Reporting customizations

Additionally, all match types can be used with these ecommerce dimensions in e.g. reporting filters and comparisons.

Google Tag Manager

New JavaScript APIs

There are multiple API endpoints available to build your own tag, client or variable templates in the server GTM.

The new readAnalyticsStorage API gives you the ability to access the current client_id and session_id from the user. Before, attempts were made to get these values from the cookies and through some kind of reverse engineering. This approach was not really stable because Google can change these cookies quite regularly.

This new API now offers solid access to this user data. With the following code, you can get the client_id from the user:

const readAnalyticsStorage = require('readAnalyticsStorage');
const analyticsStorageData = readAnalyticsStorage();

console.log(analyticsStorageData.client_id);

Further Links

This post is part of the Google Data Analytics series from datadice and explains to you every month the newest features in BigQuery, Data Studio, Google Analytics, and Google Tag Manager.

Check out our LinkedIn account to get insights into our daily working life and get important updates about BigQuery, Looker Studio, and marketing analytics.

We also started with our own YouTube channel. We talk about important DWH, BigQuery, Looker Studio, and many more topics. Check out the channel here.

If you want to learn more about how to use Google Data Studio and take it to the next level in combination with BigQuery, check our Udemy course here.

If you are looking for help to set up a modern and cost-efficient data warehouse or analytical dashboards, send us an email to hello@datadice.io, and we will schedule a call.

In this blog post, I want to summarize the new releases from the Google tools that we use daily in datadice. Therefore, I want to give an overview of the new features of BigQuery, Dataform, Looker Studio, Google Analytics, and Google Tag Manager. Furthermore, I will focus on the releases that I consider to be the most important ones, and I will also name some other changes that were made.

If you want to take a closer look, here you can find the Release Notes from BigQuery, Dataform, Looker Studio, Google Analytics & Google Tag Manager.

BigQuery

Flatten JSON Columns

In the data preparation section in BigQuery, you can now easily flatten a JSON Column. It is important that the column data type is JSON and not a STRING.

In the data preparation, you just need to click on the JSON column and then “Flatten”. If there are multiple levels in the JSON, just the first level gets flattened after the first click, but the new column can be flattened again.

It is a good feature, because flattening complex JSON structures is always time-consuming, and using this feature (with Gemini support) is a useful tool.

Comments in BigQuery

There is the opportunity to add comments to notebooks, data canvases, data preparation, and saved queries now.

The comment functionality is quite similar to the one in Google Docs, which is a good solution. You can easily switch to resolved comments as well.

However, it’s not possible to mark another person so that they receive an email. Hopefully, this will be added in an update in the future.

BigQuery ML models in UI

You can create BigQuery ML models completely in the user interface now. Just a dataset is needed, and you can create the ML model inside.

There is still the possibility to code everything, but the editor looks promising, and especially the short descriptions of the steps and choices could be really helpful.

Pipe Syntax improvement

The pipe syntax in BigQuery is another way to write SQL Code. You define every step, and the order of the statements defines the execution order. The statements are separated with the pipe symbol.

The pipe syntax got some new functionalities:

• DISTINCT
• WITH
• Names Windows

Example for DISTINCT:

(

SELECT ‘mercury’ AS item, 1 AS position

UNION ALL

SELECT ‘venus’ AS item, 2 AS position

UNION ALL

SELECT ‘venus’ AS item, 2 AS position

UNION ALL

SELECT ‘earth’ AS item, 3 AS position

)

|> DISTINCT

|> WHERE position >= 2

|> WHERE position <= 4

Data Policies on columns

Next to Policy tags, you can also use data policies to restrict access to certain columns. There is a difference in how to assign these two kinds of restrictions:

• Data Policies: BigQuery Data Policy API or DDL statements
• Policy Tags: In the BigQuery UI

Here can be found how to create, update, and delete data policies.

Dataform

No further release for Dataform.

Looker Studio

New Alignment option

There is a new chart setting available in your Looker Studio reports. With the option “Align legend with grid” you can control if the legend alignment should be based on the chart title or chart grid.

As standard behaviour, the setting is activated.

Highlight Applied Filter

The menu “View applied filter” is really useful to see which filters affect which charts.

You can see all active filters on the page, or you can select one component and see which filters are active for it.

As a new feature, you can click on one component on “Highlight” and you see a coloured border on the chart to find it more easily.

Faster responses for BigQuery data

When your report uses BigQuery data sources, you may have already recognized an improvement in the report performance.

Because the connector can use the short query optimized mode. For small requests, it returns the results immediately and does not create a BigQuery job for it.

Important: This performance boost is only available when the datasource uses Viewers or Owner Credentials.

Small Changes

There are a lot of small changes that are important to mention:

• New size for report canvas: Size A4 (portrait or landscape)

• Adding alternative texts to an image (e.g. for a Screenreader)

• Adding a report description, which is even used in report search and can be added in the edit mode of the report in the “3 dots menu” > “report details”

Google Analytics

Items data import

You can import item data to GA4, and now you can even upload item custom dimensions.

Before you upload the data, you need to create the custom dimensions in the property first.

Then, under “Data import” you can import the item data via SFTP or CSV.

Import Reddit Ads data

You can import cost data from Reddit into your GA4 property.

How it works:

• Go to the Admin Section of your GA4 Property
• In the section “Data Collection and modification” click on “Data import”
• Write a name, select “Cost Data” (nothing else is available for Reddit), select the import source “Reddit”, and click on Connect
• Then go through the authorization steps, map the fields, and the connection is set up

New default reports for focus on lead generation

When you create the GA4 Property, you need to answer, what are your business objectives.

Depending on the answer, the default reports of your property changes.

There is a change when “Generate leads” is selected as the objective.

• Two new default reports: “Lead Acquisition” and “Lead Disqualification and Loss”
• New Lead Generation audience templates

Just to mention again, the reports just show data when the right events with the right structure get sent to GA4.

Note: You can also change the business objectives later in the GA4 property settings:

Google Tag Manager

Unconsented requests in Floodlight

When you configure the Consent State in your Client GTM Container and send this information to the Server Container, the Floodlight Tags also send unconsented requests to the Floodlight server.

Further Links

This post is part of the Google Data Analytics series from datadice and explains to you every month the newest features in BigQuery, Data Studio, Google Analytics, and Google Tag Manager.

Check out our LinkedIn account to get insights into our daily working life and get important updates about BigQuery, Looker Studio, and marketing analytics.

We also started with our own YouTube channel. We talk about important DWH, BigQuery, Looker Studio, and many more topics. Check out the channel here.

If you want to learn more about how to use Google Data Studio and take it to the next level in combination with BigQuery, check our Udemy course here.

If you are looking for help to set up a modern and cost-efficient data warehouse or analytical dashboards, send us an email to hello@datadice.io, and we will schedule a call.

Latest Updates on Google Data Analytics (July 2025) was originally published in Geek Culture on Medium, where people are continuing the conversation by highlighting and responding to this story.