Swizec Teller - a geek with a hatswizec.com

Senior Mindset Book

Get promoted, earn a bigger salary, work for top companies

Senior Engineer Mindset cover
Learn more

    3 key insights that make D3.js easy to learn

    The other day somebody asked me how to learn D3.js from scratch. I quipped that it took me writing a book to really learn it. It's one hell ofa library.

    Most people don't go that far. They don't have to. Most people start with a problem, find similar examples, do some copy pasta, tweak until it works and end up with a working visualization they don't understand. You'd be surprised how few people actually understand how their D3 dataviz works.

    That was a funny tweet. It delighted 3 people, that's my bar for success.

    But the idea that there's only 1 or 2, maybe 3, key concepts you have to grok to understand every D3 example out there ... it kept me up at night. What are those 3 key insights into D3?

    1) Data manipulation vs. DOM manipulation

    All D3 examples are split into two parts: Data manipulation and DOM manipulation. First you prep your values, then you render.

    You have to go through many examples to notice what's going on. Inference learning is hard. Most beginners miss this pattern and it makes D3 look more confusing than it is.

    Let's take an example from D3's docs, a bar chart with a hover effect.

    Mike Bostock, the creator of D3, built this chart in 43 lines of code. Here they are πŸ‘‡

    Bar Chart codea

    There are two parts to this code: Data manipulation and DOM manipulation.

    Bar chart data manipulation code

    Bostock here first prepares his data:

    • some sizing variables (margin, width, height)
    • two scales to help with data-to-coordinates conversion (x, y)
    • loads his dataset (d3.tsv) and updates his scales' domains
    • uses scales to calculate attributes during DOM manipulation

    In the DOM manipulation part, he puts shapes and objects into an SVG. This is the part you then see in the browser.

    Bar chart DOM manipulation code

    DOM manipulation in D3 happens via D3 selections. They're a lot like jQuery $(something). Personally I like to do this part with React as described in my book, React+D3v4.

    Here Bostock does a few things

    • selects the <svg> node (d3.select)
    • appends a grouping <g> node (.append) with an SVG positioning attribute (translate)
    • adds a bottom axis by appending a <g>, moving it, then calling d3.axisBottom on it. D3 has built-in axis generators
    • adds a left axis using the same approach but rotating the ticks
    • appends a text label "Frequency" to the left axis
    • uses selectAll.data to make a virtual selection of .bar nodes and attach some data, then for every new data value (.enter), appends a <rect> node and gives it attributes

    That last part is where people get lost. It looks like magic. I've been using D3 for years and it still looks like magic.

    It's a declarative approach to rendering data. Works great, hard to understand. That's why I do it in React instead :)

    You can think of .enter as a loop over your data and everything chained after .enter is your loop's body. Sort of like doing data.map(d => append(rect).setManyAttributes())


    2) Scales

    Scales are D3's most versatile concept. They help you translate between two different spaces. Like, mathematical spaces.

    They're like the mathematical functions you learned about in school. A domain maps to a range using some sort of formula.

    wikipedia commons thumb d df Function color example 3 svg 440px Function color example 3 svg

    Colored shapes in the domain map to colors in the range. No formula for this one, which makes it an ordinal scale.

        let shapes = d3.scaleOrdinal()
                         .domain(['red triangle', 'orange rectangle', ...)
                         .range(['red', 'orange', ...)

    Once you have this scale, you can use it to translate from shapes to colors. shapes('red triangle') returns 'red' for example.

    Many different types of scales exist. Linear, logarithmic, quantize, etc. Any basic transformation you can think of exists. The rest you can create by writing custom scales.

    You're most often going to use scales to turn your data values into coordinates. But other usecases exist.

    3) D3 layouts

    Sure .enter.append looks like magic, but D3 layouts are the real mind=blown of the D3 ecosystem. They take your input data and return a full-featured visualization thing.

    Like a force layout using forces between nodes to place them on the screen.

    Or a circle packing layout that neatly packs circles.

    I don't know how these work internally. I've yet to try building one for others to use.

    But here's a key insight about the magic of layouts: They're the data part.

    You take a forceLayout for example and feed it your data. It returns an object with a tick event callback.

    var simulation = d3
        d3.forceLink().id(function (d) {
          return d.id
      .force("charge", d3.forceManyBody())
      .force("center", d3.forceCenter(width / 2, height / 2))

    This simulation now handles everything about rendering nodes. Changes their positions on every tick callback.

    But it is up to you to render them. A layout handles your dataviz in the abstract. You're still the one in control of rendering.

    For a force layout, you have to update the DOM on every tick of the animation. For circle packing, you render it once.

    It took me a while to get this and once I did, all the fancy looking visualizations started to make sense.


    To summarize, you need to grok 3 key insights for D3 to make sense. Once they click, a whole new world opens up.

    1. Code is split into data and DOM manipulation
    2. Scales are great and used a lot
    3. You're always in control of rendering
    Published on February 8th, 2018 in Front End, Technical

    Did you enjoy this article?

    Continue reading about 3 key insights that make D3.js easy to learn

    Semantically similar articles hand-picked by GPT-4

    Senior Mindset Book

    Get promoted, earn a bigger salary, work for top companies

    Learn more

    Have a burning question that you think I can answer? Hit me up on twitter and I'll do my best.

    Who am I and who do I help? I'm Swizec Teller and I turn coders into engineers with "Raw and honest from the heart!" writing. No bullshit. Real insights into the career and skills of a modern software engineer.

    Want to become a true senior engineer? Take ownership, have autonomy, and be a force multiplier on your team. The Senior Engineer Mindset ebook can help πŸ‘‰ swizec.com/senior-mindset. These are the shifts in mindset that unlocked my career.

    Curious about Serverless and the modern backend? Check out Serverless Handbook, for frontend engineers πŸ‘‰ ServerlessHandbook.dev

    Want to Stop copy pasting D3 examples and create data visualizations of your own? Learn how to build scalable dataviz React components your whole team can understand with React for Data Visualization

    Want to get my best emails on JavaScript, React, Serverless, Fullstack Web, or Indie Hacking? Check out swizec.com/collections

    Did someone amazing share this letter with you? Wonderful! You can sign up for my weekly letters for software engineers on their path to greatness, here: swizec.com/blog

    Want to brush up on your modern JavaScript syntax? Check out my interactive cheatsheet: es6cheatsheet.com

    By the way, just in case no one has told you it yet today: I love and appreciate you for who you are ❀️

    Created by Swizec with ❀️