Words and images have always been powerful tools for driving change throughout human history. From Greek scholars to modern scientists, people have used wax tablets, writing slates, and whiteboards to develop, explain, and discuss their ideas with their peers. This has been made possible by our innate capacity to manipulate and comprehend abstract concepts through conversation and visual cues.
Building upon this long tradition of visual conversations, we are excited to introduce our latest cutting-edge feature: Einblick Prompt. By combining the latest advancements in generative AI with Einblick's visual canvas, Einblick Prompt empowers users to construct data workflows using natural language prompts. Even complex requests like “{{remove dollar signs price, filter to older folks, and display age as a histogram}}" can be easily executed with Prompt.
You can also interact conversationally with the results, making it an intuitive and user-friendly experience.
If you can verbalize a request, Einblick Prompt will build it for you.
Teaching ChatGPT More About Data Science and Analytics
Generative AI has seen incredible recent advances and ChatGPT has upended written content, programming, and various other areas. While these chatbots are impressive and productive, end-to-end data science pipelines require a level of context and domain expertise unavailable until now.
Right now, we’re enhancing OpenAI’s large language models, such as GPT 3.5 Turbo, for data science and analytics use-cases. We’re embedding the resulting feature into our collaborative canvas, designed to display data science results and share them with your whole team. By combining the visual paradigm of Einblick’s canvas-based approach to data science with the ease and intuitiveness of natural language, we hope our users can create data insights at the speed of thought.
If you start with a higher-level Prompt, you will instantly build out an entire data science workflow incorporating Python, SQL, charts, tables, filters, and any other necessary component. Then, you can keep adding via Prompts,– or just start writing and editing code like in Jupyter notebook. If you would refine anything in Einblick, you can start chatting with Einblick Prompt directly, as if you have a project partner with you, supporting you in your data deliverables.
Why Einblick Prompt? Because normal sentences are the ultimate programming language
Why should I talk to Prompt in English, or my native Italian, rather than just writing Python? For the same reason, we don’t use Fortran or punchcards anymore. It is simpler, less prone to errors, and less painful to write. In fact, Python became data science’s default language in a large part because of Python’s simple syntax that mirrors natural language, and its reduction of boilerplate and overhead. `Hello world` is literally just `print(‘Hello World’)`.
Tomorrow’s data scientists will marvel at how we used to type 15 lines to make a simple bar chart instead of ordering a chart like a burger: “plot age histogram”
Programming languages suffer from being highly grammatical and inflexible compared to normal English sentences. No one enjoys scouring Stack Overflow posts or lengthy software documentation to find the one line that solves their current problem. Once you’ve found a solution, you have to swap out variable names, column names, etc… and hope that it works with your environment.
Einblick Prompt lets you skip all of these steps. Einblick Prompt is context-aware of metadata in your canvas, such as column, dataframe, and variable names. The context awareness enables Einblick Prompt to create functional Python and SQL tailored to your data. Einblick Prompt can also make use of our interactive components, like Charts, Tables, and Filters, to further reduce code and improve shareability.
Einblick is built on code, so you can tinker with Python and SQL directly once Prompt writes the overall flow. But alternatively, you can just open up Prompt in any Einblick cell on any canvas, and tell it how to improve that cell. You can change the colors or binning in a histogram, update the filter criteria for different analyses, or run additional hypothesis tests. Einblick Prompt becomes your personal data concierge by being there to interpret and translate your data-related requests into tangible data science outputs.
Let’s democratize data science
Data is crucial and abundant, and will only continue to grow. Currently, many organizations cannot keep up with the rate or volume of data being generated. To harness all the information that data can provide us, we don’t just need people with technical skills, we also desperately need people with industry experience and domain knowledge to be a part of the data conversation.
A product manager already knows in plain language “I want to see a graph of how session length differs by device.” Here are three ways to attack that question today:
- Instant, with no flexibility – The “dashboard of everything” to track all key metrics in static form
- Some flexibility – Buying a new tool that abstracts away code, and once the product manager learns to use its GUI, they can click away to get a specific output (within the limits of the GUI)
- Flexible, with a long turnaround time – Farm the question out to a data analyst in a shared services team
We offer a new way – just ask Einblick Prompt and get a chart back and tweak it. It’s flexible, and it’s fast. Einblick canvases are the accessible common data language for all participants. We hope that the continued development of LLMs accelerates the democratization of data science, and ensures that every person can have a voice in the data conversation.
Join the waitlist to be among the first to gain access and see what the future of data science looks like.