Observable

Observable is a platform where you can collaboratively explore, analyze, visualize, and communicate with data on the web. In this demo, you'll see how to work with your own data and discover business insights in just five minutes. https://observablehq.com/

The core Observable runtime is free and open source, while the notebook component, where people interact with data, is free but not open source.

• Large parts of Observable are open source, but not everything. In particular the editor and compiler are closed which means that you have to use the ObservableHQ web app to author notebooks. You can however embed the resulting notebook on your own website. Observable’s web editor is fantastic and is a few years (and $M in funding) ahead of Starboard Notebook's interface. However in many situations it is necessary to own the editor and not be locked in. An example such situation is when you are exploring or visualizing data that is private to your organization. This notebook showcases a plugin for Starboard that allows you to use reactive cells just like you would in Observable. Starboard's representation is simple git-friendly text files (try view source at the top of this page) - so this can be a portable alternative.

https://github.com/observablehq

A cell is a block of content. When editing a cell, the blue text box contains the cell’s source. The output (or value) of the cell is shown above... Cells can contain Markdown for styled text, JavaScript for code, and more. When editing, change the language with the editor menu in the left margin... To visualize data, create a JavaScript cell and then use Plot, Vega-Lite, D3.js, or your preferred library. If your cell returns DOM, such as SVG, it will be displayed.

How do you get data into Observable for analysis and visualization? Depending on its form and where it lives, there are a variety of ways:
inline - embedded in the notebook as code, for small amounts of data
files - attached to the notebook, for medium amounts of data (e.g., CSV, SQLite)
APIs - queried from a remote server, for programmatic access to data
databases - via an Observable database client, for accessing SQL databases... To access PostgreSQL 8+, MySQL 5, or Google BigQuery databases, you can use an Observable database client. For databases on private networks, use our self-hosted Observable database proxy.

• there's a "Datasette Client - with SQL Cells"

visualization, jupyter-like "notebook" structure: Computational Medium?

CTO is Mike Bostock