Automatically generate beautiful visualizations from your data

And other bad ideas

Ian Johnson
5 min readJan 12, 2016

I work in data visualization, a loosely defined and rapidly evolving field that is generally about taking data and turning it into something people can understand. There are a lot of different tools and disciplines involved in doing so, and it is impacting almost every field of study, industry and business. I believe it is essentially a new medium for communication, and we are still very much in the formative stages of its development.

When people ask me what I do, I tell them something like “I make data visualizations for the web”, with more or less detail depending on how tech savvy the person is. I can’t count the number of times I’ve been told with enthusiasm:

I’d love to have something where I can put in my data and it gives me back a beautiful visualization!

I’ve heard it so often I’m never surprised when I meet someone who is starting a startup to “make data visualization easy” or some combination of the words “one-click”, “magic”, “anyone can do it” and “data viz(sic)”. When an idea is obvious to everyone, including non-experts, it’s a pretty big warning sign that it’s not a good one.

96% of datavis startups use a screenshot of the d3.js gallery for their marketing materials

I want to take a deeper look at why this pursuit of automation is misguided, and in the process hope to point out potentially more fruitful paths. I intend to do this by looking at how other communication mediums have come about via technology, what the authorship tools look like and how they evolved. We will start with the most recent medium and go back in time, getting deeper into the essence of augmenting human communication with technology.


I’d love to have something where i can drag-and-drop my iPhone footage and it spits out Tarantino quality cuts!

Arranging moving pictures to tell a story is a pretty big deal in our society; from movies to TV, from news to education we’ve whole-heartedly adopted this medium for communicating complex ideas.

The input to movies is video, something that in its raw form is readily understandable by everyone. This input is then edited and post-processed into its final form ready for consumption. The tools for creating the input are ubiquitous, and the number of videos on YouTube alone indicates how many people are willing to endure the editing process in an attempt to express themselves and their ideas.

One thing you don’t hear very often is the wish for a magical tool that magically turns people’s video into an award winning feature. Sure, there is iMovie and a bunch of pre-canned effects and filters, but a quick survey of YouTube will show that they don’t provide enough magic to make every video into something worth watching.

What we have instead is a stable of professional tools that occupy a spectrum of specialized features. Editing tools like Premiere, Final Cut Pro and AVID, special effects tools like After Effects and Cinema 4D or animation tools like Lightwave. Most of these have complex user interfaces which take training and years of experience to master. Some of them have entry-level versions to help beginners get started, but the general idea is that if you want to make high-quality moving pictures you can expect a learning curve and to work with a team of dedicated craftsmen (or craftswomen).

Personally, I believe data visualization has the potential to become as important to society as film, especially as the explosion of computer applications power more ever more complex processes. I’m not sure what the future holds, but I’m pretty sure it’s more powerful and complex tools rather than simpler and easy to use.

Now Let’s examine another communication form that had an essence before technology came along and changed everything about how it’s done.


I’d love to have something where I can sing into it and sound like Beyoncé.

I know what you’re thinking: Auto-tune. How many Beyoncé’s has Auto-tune produced? None. Auto-tune and digital editing with professional tools like the aptly named ProTools can turn the human voice into almost any sound we want now, but it can’t make bad music into good music. The closest it can come is making bad sounds into “good” sounds, but sound is only one component in music. Other components can include emotion, cultural relevance, lyricism and the catch-all uncomputable factor: creativity.

There are many ways to apply good aesthetics to a data visualization, there are even some rules that you can choose to follow to accommodate the way our brains perceive shape and color. A computer program could be encoded with a designers aesthetics to consistently produce “pretty” artifacts, but just like music that wouldn’t be enough to generate a compelling visualization.

An effective visualization much like an effective song must effect something in the consumer, a feeling, an understanding, it must communicate from one human to another. There simply isn’t an algorithm for that yet, and once there is we will have much bigger things to worry about.


I’d love to have something where I give it my ideas and it spits out elegantly articulated paragraphs

For my final argument I give you clippy. Q.E.D.

Everyone in the developed world is familiar with word processing software by now, and its impact on our ability to communicate with the written word is so obvious it largely goes unstated. We’ve been using technology to increase our effectiveness at writing for centuries, from the ball-point pen, to the movable press, to the typewriter to Microsoft Word we haven’t stopped innovating. What we have stopped doing is wishing for a magic genie that would turn our messy thoughts into coherent writing, for that we just hire an editor.

Just like with writing, there is a lot of thought that needs to go into communicating with data visualization. It is not enough to just have a collection of notes or a spreadsheet of data, the author must analyze the subject, identify the interesting findings and find a way to present them in a way that is amenable to her audience. There is no shortcutting this without sacrificing the quality and effectiveness of your communication.

Unlike writing, the process and techniques for working in the data visualization medium are new and require some technical knowledge and familiarity. The opaqueness of the process for most people makes it easier to succumb to hyperbolic claims and the allures magic elixirs, so I will end with a useful technique for spotting a bullshit idea.

This trick can be used by anyone regardless of technical knowledge, just replace occurrences of the word data with idea and visualization (or dashboard) with article. You may also choose to replace developers with writers.

From uploading your idea to using it to create articles, there is no step in between to complicate matters (and no need to hire a team of writers to get started). Just upload your raw idea to start using it to create an article.

That’s all folks

If you like this article, follow me as I continue to turn ideas into articles and data into visualizations until one of these deep learning neural networks startups proves me wrong.



Ian Johnson

pixel flipper. Data Vis Developer @ObservableHQ. formerly @Google @lever