Dataslope

Courses and coding playgrounds that run entirely in your browser. No install, no setup, no sign-up, no paywall.

dataslope.com | Courses | Interview Prep | Playgrounds | Pricing

What this is

Dataslope is a site for learning programming, data work, and databases. It has two halves, and they share the same engine:

• Courses. Written lessons with runnable code, auto-graded challenges, and quizzes built into the page you are reading.
• Playgrounds. Full editors for fourteen languages and database engines, one click from the nav bar.

Every runtime is compiled to WebAssembly and executes on your own machine. There is no queue, no container to provision, and no round trip to a build server: open a lesson and a real Python, C++, or PostgreSQL runtime boots next to the text.

Today the site holds 32 courses and six interview-prep tracks, roughly 900 pages carrying 3,800+ runnable code blocks, 800+ auto-graded challenges, and 3,200+ multiple-choice questions.

Nothing is gated, and nothing will be

Every lesson, exercise, challenge, quiz, and playground is readable and runnable with no account. Not a trial, not a preview, not the first three chapters. All of it. Anonymous readers get the same statically served pages a signed-in member does.

This is a commitment, not a current state of affairs. From the Terms:

All learning content available today is completely free; we may introduce additional or optional paid features in the future, but doing so will not require payment to access the content that is currently free.

An account is optional and free. It adds cloud saves for your workspaces and manageable share links. That is the entire difference. There is no paid tier today, and the learning content is not what a paid tier would ever be for. The details, including the storage limits and the capacity caveats, are spelled out on the pricing page.

Unlimited code executions

Because the runtimes are WebAssembly running in your tab, there is nothing to meter. Run a snippet once or ten thousand times; it costs the site nothing, so there is no execution quota, no daily limit, and no "you have 3 runs left today" banner. Your code stays in the browser unless you explicitly save or share it.

The practical upshot: you can sit on one lesson and keep changing a number until the output finally makes sense. That is usually where the learning happens.

Languages and engines

The same fourteen runtimes power the playgrounds, the lesson code blocks, and the challenges.

	Language	Version	Runs on
	Python	3.14	Pyodide (CPython compiled to WASM), with NumPy, pandas, Plotly, Matplotlib
	R	4.6	WebR, with ggplot2, dplyr and friends
	JavaScript	ES2023+	The browser's own engine
	TypeScript	5.9	In-browser transpilation, with a real language service
	PHP	8.4	php-wasm
	C	C17	Clang, compiled to WASM in the tab
	C++	C++20	Clang, compiled to WASM in the tab
	Java	8	CheerpJ (OpenJDK)
	C#	14 on .NET 10	Roslyn on .NET WebAssembly
	HTML / CSS	HTML5, CSS3	Live preview of your own markup
	React	19.3	JSX bundled in the browser
	PostgreSQL	18	PGlite
	SQLite	3.53	The official SQLite WASM build
	DuckDB	1.32	DuckDB-Wasm

Playgrounds

Three kinds, depending on what you are doing:

• Code editors for the general-purpose languages: file tabs, a real multi-file project, autocomplete, formatting, and an output pane that understands text, data frames, charts, and figures.
• Web sandboxes for HTML/CSS/JS and React, with a live preview beside the editor.
• A SQL workbench with a schema browser, query tabs, an editable result grid, EXPLAIN, and sample databases already loaded.

General-purpose code editors

Multi-file projects with tabs, not a single textarea. Run the file you are on, or pick another entry point from the Run menu.

C++. Clang compiles to WebAssembly in the tab, and the binary never leaves your machine.

Java. OpenJDK by way of CheerpJ.

C#. Roslyn on the .NET WebAssembly runtime.

C. The same Clang toolchain, C17.

JavaScript. The browser's own engine, with Node-style module resolution.

TypeScript. Transpiled in the browser, backed by a real language service.

PHP. php-wasm, 8.4.

Web and React sandboxes

HTML, CSS, and JavaScript across separate files with a live preview, and React with JSX bundled in the browser as you type.

SQL workbench

Not a query box. A schema tree with tables, views, indexes and triggers; multiple query tabs; a sortable, filterable, editable result grid; EXPLAIN; import and export. Three engines, each with sample databases already loaded.

Inside a course

Courses are prose first, because the explanation matters more than the widget, but the page is live wherever being live helps.

Each course opens on a welcome page with its outline in the sidebar, and every page is reachable without an account.

Lessons are illustrated. Every course and most lessons carry their own artwork, diagrams, and figures rather than stock screenshots.

Charts and diagrams are drawn for the point being made, not borrowed from somewhere else.

Code blocks run in place. Edit the snippet, press Run, see the output under it, in the lesson rather than in a separate tab.

Challenges are auto-graded against real test cases, and some span multiple files. There is a Solution button when you want it.

Quizzes explain themselves. Each choice comes with a reason, right or wrong.

Course pages also have full-text search across the whole site and per-lesson Markdown export.

Interview prep

Six role tracks (data analyst, data scientist, data engineer, analytics engineer, machine learning engineer, backend engineer), built the same way as the courses, so every question is one you actually run rather than one you read the answer to.

Under the hood

Next.js on Cloudflare Workers, with the lessons prerendered and served statically. Language runtimes are WebAssembly, loaded on demand and cached by the browser. Search is a SQLite FTS5 index on D1. Accounts, when you want one, are Better Auth on D1.

The engineering detail (deployment, caching, search indexing, auth, the admin surface) lives in DEVELOPMENT.md. Repository conventions and content-authoring rules live in AGENTS.md.

Feedback

Bug reports, corrections to lessons, and requests for languages or topics are all welcome in issues and discussions. Corrections to content are especially useful; a wrong explanation is worse than a missing one.

License

Dataslope is available under more than one license:

• Source code: MIT License. The MIT license covers Dataslope's own code only, not the learning content, and not the third-party software, runtimes, datasets, fonts, or artwork the project depends on or loads at runtime, each of which keeps its own terms.
• Learning content: Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0). Attribution required, non-commercial use only. This covers the courses, lessons, exercises, and quizzes under content/ and the illustrative teaching materials made for them: the lesson and course illustrations in public/images/, and the figures, diagrams, and charts that accompany the lessons (including the ones rendered at build time from charts/). The chart scripts are code under MIT; the drawings they produce are content. See LICENSE-CONTENT for the exact boundary, including what is excluded (brand assets, third-party material).

Third-party software and language runtimes retain their own licenses; see THIRD-PARTY-NOTICES.md. In particular, the Java runtime CheerpJ (Leaning Technologies) is proprietary, used here under its free Community Edition and loaded from Leaning Technologies' CDN.