Name: Hex Magic
Uploaded: 2025-02-19T10:22:43.600Z
Duration: 2 min 33 s
Description: An AI-powered analytics assistant that streamlines data analysis by writing queries, building charts, and fixing bugs.

hex-magic

An AI-powered analytics assistant that streamlines data analysis by writing queries, building charts, and fixing bugs.

Hex Magic

Magic is a tool that puts the power of hex into a single prompt allowing you to answer questions faster using natural language. With magic you can write or edit SQL queries, python code, create charts and automatically correct mistakes using AI. Wherever you are in your project you can get started with magic by selecting the use of magic button or hitting command shift M. This will

Open up the prompt bar which is your portal into everything magic is capable of. Magic has a deep understanding of your data models, table schemas, and query conventions to ensure that prompts are executed correctly. Notice that as I'm prompting magic, I can even use the at symbol to directly reference a table from my database. Once magic has finished thinking, we can choose to accept or

Discard its generated suggestion. Accepting these changes, we'll switch the prompt bar into edit mode, which we can use to fix, explain, or change existing code. Let's edit the cells to filter out any spicy menu items. 

Magic is not limited to the confines of a single cell. It can generate new cells just like you would while working in a notebook. We can see this in action by.

Selecting add new cells just below the prompt bar and then entering in our prompt. Magic understands the context of your project allowing it to refer to existing data frames or variables and even create new ones as needed. In the integration with chart cells enables magic to create consistent on-brand visualizations that supports build interactivity like cross filtering and.

Drilling magic also respects your data privacy only ingesting high level details such as schema and column names not the actual records in your database. While this ensures data privacy, it means magic doesn't know exactly what your data looks like leading to a few common situations where it struggles. For example, here it doesn't know the format of values in this column so it.

Doesn't know that completed shouldn't be capitalized to help magic out. We can use a data manager by default. Magic automatically ingest metadata from DBT models and warehouses to help guide its suggestions. The data manager allows you to give magic extra context such as the format of column values or how certain tables relate to one another. This can be done at the database level.

The schema level the table level all the way down to the individual columns. At each step of the way you can tell magic what it should pay attention to, what it should ignore and what it should prioritize. The best part of working with magic is that it's deeply integrated into your hex project. So you can work manually when you want to and rely on Magic whenever it's helpful.

When you're done working you can easily turn your analysis into a data app that can be shared with the URL. Now anyone on your team can explore the insights that you and Magic come up with.

Note, that there may be slight differences in text and video. Watch video for the original content.

Hex Magic Video