Data Pond

Sometimes you don't have a lake, but a little pond of data sitting around in your computer and you would like to explore it.

Data Pond is a little app for just that:

• Import a folder full of CSV files
• Turns each of them into an in-memory table (powered by Duckdb) and show them on a sidebar with their columns and inferred data types
• Let you run SQL queries over them, including creating new tables that will show up in the sidebar
• Simple plotting of results

This is more of a toy application, that I made as way to practice the approach described in the GOOS book, but I still find it neat and have used it ocasionally. I see it as just a bit less bureaucratic than firing up a Jupyter notebook and fiddling with data there, that I would rather open it to look at some files.

How to

To run it or develop it you need:

• Python 3.12*
• make

*present on your PATH as python3, or you can redefine the SYSTEM_PYTHON env var to point to your installation before running make .venv

• make sync creates a venv under .venv, and install dependencies.
• make test runs tests
• make run runs the app

I use uv as a drop in replacement for pip-tools, but it's installed automatically in the venv. I will probably switch to pure uv soon, now that the dependency locking mechanism is in place.