Dataviz.Garden

Powered by 🌱Roam Garden

DataViz – to Code, or NOT to Code?

That is the question! And the answer is: it depends.

No-Code Way

Lots of people create useful visualisations and never code.

Tableau is one tool that can get you far.

There's also a number of fairly new, exciting no-code tools like Flourish, DataWrapper or RawGraphs.

Code Tao

More complex, interactive and web-based data visualisations tend to use solutions which involve coding. These are often created by teams consisting of data developers, researchers and designers.

That being said a lot of people code beautiful dataviz pieces on their own. One of best places is ObservableHQ with its "live-coding" environment.

Web-based, visual-oriented pieces are mostly programmed in JavaScript. Not pure ("Vanilla") JS but leveraging one of the existing frameworks and libraries.

There's a distinction to be made between high-level, general-purpose libraries...

There's plenty to choose from these days.

...and low-level, viz-it-as-you-like, complex libraries. These take longer time and more skill to work with. d3 is the dominant one (also often leveraged by the high-level libraries under the hood).

More data-science kind of visualisations is usually processed in R or Python. Both are good options with rich dataviz libraries available.

Examples

Moritz Stefaner is someone who both designs and often codes data visualisations.

He discusses his favourite JavaScript framework (Svelte) in 163th episode β†—οΈŽ of his podcast Data Stories.

Interestingly, David McCandless doesn't code when creating his famous infographics, despite having worked as a programmer early in his career.

You can see his graphics on Information is Beautiful web. The pieces often don't need code. When they're interactive, David hires a developer to do the coding.