Notebooks for Everyone

How Noteable’s ChatGPT Plugin Empowers Everyone to Do Data-Driven Analysis

Elijah Meeks
Noteable
6 min readMay 11, 2023

--

The capabilities of OpenAI’s ChatGPT have grown even stronger with the release of Noteable’s notebook plugin. Now, users from all backgrounds can use natural language prompts to create computational notebooks with exploratory analysis, visualization, machine learning, and data manipulation.

All about our new plugin in just under a minute

Noteable’s mission is to “Enable everyone to work with data the way they want.” To achieve that goal, we built a modern notebook that kept to Jupyter standards and added convenient git integration, secure connections to data sources, managed secrets, commenting, and even a data visualization experience rivaling popular BI tools. All this enables users of all backgrounds and skill levels to work with data the way they want. We envision AI as allowing us to extend our mission beyond traditional data roles like data scientists, data engineers, and analysts to anyone who wants to deliver data-driven insights.

How it works

In the past, you needed technical skills to work with notebooks. Beyond knowing how to code, you also needed system administration experience to set up the environment, install libraries, and manage your project with little UI support. Modern data notebooks improved this process by providing better interfaces, integrations and environments so that you can focus on working with data and playing with ideas.

Noteable’s ChatGPT plugin changes all of that. By describing the data you want to work with, the analysis you need, the techniques you want to explore and how you want all that brought together in a literate programming document, you’ll receive an entire notebook as a result. And that notebook isn’t just some static report, it’s an interactive and collaborative document that you can use with your stakeholders and peers to develop your work further or schedule it to facilitate periodic discussions. Imagine how you can just prompt chatGPT to build a notebook to have all your product data analysis to guide your MBRs or analyze your user traffic or visualize your revenue streams. That’s possible now!

Let’s not lose sight of just what a big deal this is. We’re at a point when all the barriers to doing data-driven work are coming down. What started with convenient UIs, better APIs and stackoverflow has been amplified by the integration of ChatGPT giving you a data-driven content generator for everyone to transform ideas into impact.

How to get started

Why notebooks are the perfect document for AI work

Having a universal document for collaboration– a notebook–combines the best of ChatGPT assistance, such as code generation, error explanation, and intelligent prose, with tried and true data science workflows. Notebooks have always been the perfect medium for blending text, code and visualization; using natural language to create them aligns perfectly with their conversational nature. The clear separation of text and code cells, along with the rich history and culture surrounding notebooks, creates a solid foundation for collaborative work between people and AI. And after you’re done creating the notebook with the plugin, it’s there to be extended, expanded, forked and commented on.

We must not forget the importance of working together to make sense of data, instead of relying solely on AI-generated answers. Automation didn’t just begin with ChatGPT–it’s a core principle behind UIs that expose complex tasks with simple controls. We shouldn’t throw that away in the hopes that all we need is a quick answer from an AI.

That’s particularly true now, when the answers given back by the AI in the realm of data work are rather simple. But it will still be important when the answers get better–perhaps even more important. We need to make meaning together, and not rely on a tool to tell us answers. What we need from these tools is for them to help us bring more people into a conversation about the meaning of data, not to replace that whole conversation with a relationship with an oracle.

Assisting Data Work

We’ve seen AI integrated into data tools already. Noteable itself leverages the OpenAI API to give the user feedback about errors and %%assist prompting. These features help data scientists, analysts and data engineers do their work better. But this plugin does more than that–it opens up the ability to do data science, analysis and data engineering to anyone who can write a prompt. When it’s finished, you have a document available for future work and collaboration. This opens up so many more possibilities when it comes to how we envision working with data in teams. Despite our excitement with these new features, we know that Noteable has a responsibility to consider and foreground the ethical concerns around this new work and we’ve been thinking carefully about this topic.

A world where everyone can code

The most striking feature of this plugin is its ability to let domain experts who don’t happen to have experience coding or doing data analysis to now create data-driven documents. In the past, their contributions relied on and were mediated by the skills and availability of collaborators. Now they can create their own documents and work within them or bring in other experts to improve the document.

Communicate with everyone using translation

One of the chief concerns we have when it comes to natural language interfaces is that they preference English speakers. But it’s also true that these tools, with their ability to translate into other languages, can now create reports and support non-English speakers. The availability of tools like Noteable’s notebook plugin will require all of us to think about how we can better empower people who have traditionally experienced these barriers to access.

Use the best from different coding languages

Even when they can code, not everyone can be a full-stack developer. To do work in the modern data stack, you can’t get by on python alone. Understanding how to write a query in a particular dialect of SQL or how to write configuration for the countless tools and libraries necessary for data science is necessary to efficiently work with data systems. This plugin excels at teaching these skills by providing documented samples that solve 90% of most data problems.

Everyone receives superpowers of computation, communication, and collaboration

Our plugin isn’t just to enable access to notebooks for those who traditionally haven’t used them. It also serves as a tool for creating effective projects and increasing the number of options available to experienced notebook users.

A deep dive into how you can use our plugin to get started on your data-driven journey

Where will it take you?

Three years ago, we began building a notebook platform for everyone to tackle the exponential growth of data and increasing complexity in data toolchains. Today, we’re proud to offer a third-generation notebook platform that integrates AI-driven support during notebook creation and use, enabling data professionals and executives alike to make sound decisions and analyze data with ease.

Noteable combines the best elements of Business Intelligence (BI) and Artificial Intelligence (AI) to provide everyone in your organization with a knowledge generation solution. It unlocks greater confidence in your data workflow, allowing you to make informed business decisions. We’ve developed a platform that seamlessly blends modern data stack features with capable agents to assist you throughout the notebook experience. With our ChatGPT plugin, we’re lowering the barrier to entry, enabling any user to get started on their data project with rich description and code.

Even if you’ve played with the other features that ChatGPT exposes for creating code or graphics, you should try the Noteable plugin. The ability to create interactive, living documents with code and text and visualization that you can return to with your stakeholders and team truly is a new experience. That’s what makes a notebook great: access to databases, environment variables, package installs, GPUs, being able to edit the code & text, share the resulting notebook with others and collaborate via comments and published notebooks.

--

--

Elijah Meeks
Noteable

Principal Engineer at Confluent. Formerly Noteable, Apple, Netflix, Stanford. Wrote D3.js in Action, Semiotic. Data Visualization Society Board Member.