• Skip to main content
  • Skip to primary sidebar
  • Skip to footer

eagereyes

Visualization and Visual Communication

  • Explore
    • Starter Pack
    • Blog Calendar
    • Blogroll
    • eagereyesTV YouTube Videos
  • Practical
    • Basics
    • Pie Charts
    • Techniques
    • Book Reviews
    • Journalism
  • Academic
    • Speaking Mistakes
    • Acceptance Rates
    • Papers
    • Conference Reports
    • Lists of Influences
    • Criticism
    • Peer Review
  • Admin
    • About
    • Contact
    • License
Encoding vs. Decoding

Robert Kosara / February 20, 2017

Encoding vs. Decoding

Visualization techniques encode data into visual shapes and colors. We assume that what the user of a visualization does is decode those values, but things aren’t that simple.

Encoding

When a program draws a bar chart, it calculates the length of the bars from the numbers it’s supposed to represent. When it draws a pie chart, it calculates angles. When it draws a scatterplot, it looks at two numbers for each data point and turns those into coordinates to draw a shape.

We understand the encoding part very well. There’s nothing mysterious about how a chart comes about, it’s a mechanical process.

This is also where we have much of the theory of visualization, such as it is. Bertin’s retinal variables, despite their name, are all about encoding. Likewise, Wilkinson’s Grammar of Graphics has a formalism for many different ways of encoding numbers.

Data properties, like whether a data field or column is numerical vs. categorical, whether there’s a meaningful zero, etc., are all about encoding. This is the stuff that lives inside the machine, and that we can formalize very easily.

Decoding

When it comes to decoding, things get a lot messier. What do we decode? We like to assume that decoding just reverses the encoding: we read the values from the visualization. But not only don’t we do that, we do many other things that are surprisingly poorly understood.

In a bar chart, we rarely look at individual bars. Instead, we compare them to each other. We also look at the shape of the plot. Which is why being able to sort a chart is incredibly important (charts are from this posting on problems with election maps).

In a pie chart, we presumably compare a slice to the whole. Except we also compare slices to each other, of course. And that thing about the angle? Well, we’re apparently not actually reading it that way.

It gets even more interesting when we look at complex charts like scatterplots, parallel coordinates, etc. In a scatterplot, nobody compares two data points along two axes. Instead, we look at the overall shape. That gives us an idea of correlation (which we’re surprisingly good at estimating), clusters and density, and outliers. Those are much more interesting than merely looking at data values.

The Value of Visualization

If visualization were about decoding values from charts, things like aspect ratio would not matter; the number of bars between the two you’re comparing would not matter; the orientation of slices in a pie chart would not matter.

If it were about decoding values, we would not be getting anywhere as much out of visualization as we do. We would not be getting correlation or clusters or outliers from a scatterplot. There would be little point in drawing pictures from data at all.

What makes visualization powerful is our ability to go beyond the mere decoding of values from a chart. That makes it interesting, but it also makes it complicated. So far, we have focused our understanding largely on the encoding side of visualization. We need to learn much more about the complex and powerful decoding side.

Filed Under: Basics

Robert Kosara is Data Visualization Developer at Observable. Before that, he was Research Scientist at Tableau Software (2012–2022) and Associate Professor of Computer Science (2005–2012). His research focus is the communication of data using visualization. In addition to blogging, Robert also runs and tweets. Read More…

Reader Interactions

Comments

  1. Bilal says

    February 20, 2017 at 9:47 pm

    Very good point, thank you Robert

    Reply
  2. Ryan Shaw (@rybesh) says

    February 21, 2017 at 9:46 am

    One good way to start understanding the value of visualization better might be to choose a new metaphor: https://goo.gl/zhZKgh

    Reply
    • Robert Kosara says

      February 21, 2017 at 11:31 am

      Ryan, that link doesn’t appear to be working. I’m just getting an Access Denied error.

      Reply
      • Alex Cookson says

        February 21, 2017 at 11:50 am

        I had the same issue. Here’s a working link:

        http://www.academia.edu/6290150/The_Conduit_Metaphor_A_Case_of_Frame_Conflict_in_Our_Language_about_Languag

        Reply
        • Stephen Hampshire says

          March 20, 2017 at 4:18 am

          Excellent blog post, and an excellent article too. I particularly like “Human communication will almost always go astray unless real energy is expended” p. 295

          Reply
      • Alex Cookson says

        February 21, 2017 at 11:55 am

        I’ll also follow up with a link to the Wikipedia article. As a layman, I found the article difficult to follow.

        https://en.wikipedia.org/wiki/Conduit_metaphor

        Reply
  3. Michael Kramer says

    March 2, 2017 at 8:40 am

    This brought to mind the famous essay by Stuart Hall, https://faculty.georgetown.edu/irvinem/theory/SH-Encoding-Decoding.pdf
    https://en.wikipedia.org/wiki/Encoding/decoding_model_of_communication

    Reply

Leave a Reply to Bilal Cancel reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Primary Sidebar

More Basics Articles

  • Row-Level Thinking vs. Cube Thinking
  • Spreadsheet Thinking vs. Database Thinking
  • Visualization Research, Part I: Engineering
  • Putting Data Into Context
  • What Means Mean

Recently Popular

  • The US ZIPScribble Map
  • New, Improved Traveling Presidential Candidate Map
  • Chart Junk Considered Useful After All
  • Data: Continuous vs. Categorical
  • The Travelling Presidential Candidate Map
  • New video: Gauges for Data Visualization, The NY Times Election Needle, and Circular Bar Charts
  • The Simple Way to Scrape an HTML Table: Google Docs
  • Facebook
  • GitHub
  • LinkedIn
  • RSS
  • Twitter
  • YouTube

Subscribe via Email

Footer

  • About
  • Contact
  • License

Copyright © 2006–2022 Robert Kosara · All original materials are available under CC-BY-SA

 

Loading Comments...