For years, Looker has been a trusted platform for building reliable, governed data experiences. You’ve used Looker to build intelligent, flexible dashboards and Explores that empower you and your colleagues to make data-driven decisions.
Today, we’re excited to introduce Code Interpreter, a powerful new capability within Conversational Analytics that brings the power of Gemini to your Looker data. Code Interpreter helps you answer complicated questions (e.g., “Which of my customers are likely to churn?”) that used to involve queuing up in your data science team’s backlog and possibly waiting days or weeks for an answer.
When the old way doesn’t cut it
We consistently hear three major challenges from teams who want to use their enterprise data more actively:
- Analysis takes too long: Answering complex, ad-hoc questions can take days or weeks.
- Too few people have access: Advanced analysis is often limited to a small number of experts with SQL or Python skills.
- Advanced insights are expensive: Predicting sales, estimating customer churn, and running other types of machine-learning-based analysis require specialized data science resources that are in high demand.
By bringing Conversational Analytics to Looker, the barriers to getting informed answers to drive decisions practically melt away, as everyone can talk to their data using natural language. With the debut of Code Interpreter, you can now take on data science and make these one-time challenges seem like relics of the recent past.
Code Interpreter: Your on-demand data scientist
Code Interpreter is a new mode within the Conversational Analytics experience. When you enable it, you unlock new possibilities beyond the calculations defined in your Looker Explores. Instead, you unlock the full flexibility of Python to perform advanced analysis on the fly.
Think of Code Interpreter as an on-call data scientist available directly in Looker, ready to answer your most complex questions in seconds at any time.
Nothing typical about it
Code Interpreter excels at questions that go beyond the scope of the standard BI query, helping you reimagine what’s possible when you can truly engage with your data.
For example, here are some new tricks you can unveil:
-
Perform Complex Period-over-Period Analysis You no longer need complex table calculations or custom LookML to compare periods. You can now ask directly for quarter-over-quarter or year-over-year analysis, even with custom offsets.
Ask: “What were our quarterly sales in 2023 and 2024, and what was the quarter-over-quarter growth?”
-
Uncover Key Business Drivers Go beyond simply reporting what happened and start to understand why. Code Interpreter can analyze multiple dimensions at once to summarize the top factors influencing a key metric.
Ask: “What were the key drivers of sales in my data?”
-
Conduct Advanced Customer Segmentation This is where the power of Python really shines. You can use statistical techniques like k-means clustering to segment your customers into meaningful groups without writing a single line of code yourself or going deep into the machine learning world.
Ask: “Perform a customer segmentation analysis to group my customers into three cohorts.”
How it works: Transparency and trust
We know that in the world of data, trust is everything. An answer is useless if you can’t verify how it was calculated. That’s why Code Interpreter shows its work. For every answer generated, you can expand a “How is this calculated?” section to see the exact Python code that was run. It’s not a black box.
This transparency is key, because the only thing scarier than a snake in your boot is a number you can’t back up in your board deck.
This is also why Code Interpreter is a distinct mode you choose to enter. The core Conversational Analytics experience strictly respects the metrics defined in your LookML, preserving your single source of truth. Code Interpreter gives you Python’s flexibility for deeper discovery, with the understanding that you, the user, can scrutinize the results.
Getting access to Code Interpreter
Code Interpreter is available in Looker Studio Pro, Looker (original) 25.8+, and Looker (core) 25.10+.
Your admin can enable it in the Admin panel under Gemini in Looker. Once enabled, users will see an “Advanced analytics” toggle in the Conversational Analytics window.
Feedback is a gift!
We are eager to see you give Code Interpreter a try and share early feedback to shape our roadmap. You can share feedback directly with our team here: conversational-analytics-feedback@google.com. And of course, you can turn to the Looker Community forums to share your thoughts, questions, and success stories.