Matrices helps you quickly explore large datasets (100k rows) without needing SQL or code. It was borne out of my frustration with the current state of data visualization tools. They tend to be slow and require endless customization.

Easily create and share graphs like this. It's interactive, give it a spin.

Matrices takes a different approach. Instead of giving you endless customization, it's a general purpose 'data swiss-army-knife'.

  • want to filter down some data?
  • want to generate a nice looking graph for a presentation?
  • want to get a sense of a dataset that you've never seen?
  • want to share a graph with a colleague?

Matrices is great for all of those things. It's supposed to feel really lightweight, like "gist for data".

Find the data you need with just a few clicks. 50k rows in 1 second.

Have fun, and let me know what you think.


How do I get started?

There's a couple of core things to know about Matrices:

  1. when you load in data, it should be a 'well formatted' table. That means it has a header row.
  2. once you've loaded your data, you can take 'actions' on it: filter, group, join, etc. These actions 'stack' on top of each other, each one will give you an output table
  3. you can graph any output table by switching to 'graphing mode'.

To get started, simply upload a CSV or paste in your data from excel or google sheets (but remember, it should have a header row!).

What isn't it good for?

  • if you feel most comfortable with code and doing deep data analysis, you'll probably be frustrated with the limited functionality
  • if you have a large dataset (10mb+) with 'wide rows', Matrices won't work super well
  • if you have a dataset with a lot of free-form text like survey responses. Matrices is designed for structured data

Where is my data stored?

Today, all of your data is stored locally. It doesn't leave your computer, other than a small sample of the rows that hits the OpenAI APIs.

That changes when you hit 'publish', to either publish a dataset or a graph. Then your data will be uploaded and stored on our servers so others can access it via the URL.

For analytics purposes, I do store some metadata about your usage of the app. Namely, things like column names and database metadata. I don't store any of your data, though.

Can you add feature X or fix bug Y?

I'm always looking for new feature ideas. Shoot me an email: [email protected].

If you're sending a bug report, remember that I don't see any of your data, so if there is specific functionality that is causing a bug, you'll need to either send me the dataset or describe it.

How does the product make money?

It doesn't yet. My goal is to help people all over the world explore data. It costs me some money to run the app, but I'm happy to do it. That may have to change depending on my infrastructure costs.

I've been mulling over including a pro plan which would do a lot more. If you're interested in this, please reach out: [email protected].

What's the roadmap?

There's a lot of other stuff that I think would be useful here...

  • excel-like formulae
  • additional data sources from databases and warehouses
  • additional commands like 'pivot' and 'unpivot', joins
  • additional graph types
  • better sharing + discovery features

But I'm just one guy. We'll see where it goes.

How do I get in touch?

Shoot me an email: [email protected]. Feature requests, bug reports, or just to say hi.

Who is behind this?

Hi there, I'm Calvin. I like building things that help people. I've always been fascinated by using data to get insights, and annoyed at how much python and SQL have stopped regular people from doing that.

I'm hoping Matrices can help.

This project is also a playground for testing out new AI features and seeing just how far a solo dev can go.


Thanks to Fabian at Tonik for the designs, and to Marissa Montgomery for help building this out.