Building custom

data visualizations

Shirley Wu

@sxywu

shirley wu

 

independent creator of

data visualizations

(sxywu.com)

 

co-organizer d3.unconf

instructor fem

My Fem workshops

Coming from React:

 

Data Visualization

for React Developers

Intro to D3

Building Custom

Data Visualizations

From elsewhere:

 

Intro to D3

Building Custom

Data Visualizations

(maybe)

Data Visualization

for React Developers

custom data visualizations can be categorized into two broad categories:

EXPOSITORY

VS.

EXPLORATORY

expository

- static dataset

- explore data for story

- communicate story to audience

exploratory

- dynamic dataset

- interview stakeholders

- build tool for stakeholders to explore the data 

examples:

New York Times, The Pudding, The Washington Post, etc.

examples:

scientific visualizations, internal business tools at Netflix, Uber, Airbnb, etc.

expository

Data: top 10 blockbusters every year for the last two decades

 

Goal: come up with a design and implement it together

(Yay participation!)

 

(The raw data)

Data exploration

with Observables

and Vega-Lite

Design with

Marks, Channels

and Gestalt Laws

Code with

SVG paths

and D3 shapes, layouts

Finish with

annotations, axes, legends

data exploration

  1. List data attributes
  2. Ask questions
  3. Explore the data

 

We will use an Observable notebook for this.

data exploration:
How to use observables

To load external libraries: d3 = require('d3')
To write text: md ` ... `
To set global variable: globalVar = ...
To write a block of javascript code: {
  ...
}
To create an SVG element: {
  const svg = DOM.svg(width, height)
  ...
  return svg;
}

data exploration:
data types

  • Categorical (movie genres)
  • Ordinal (t-shirt sizes)
  • Quantitative (ratings/scores)
  • Temporal (dates)
  • Spatial (cities)

data exploration

List all the attributes,

ask all the questions!

(Notebook)

data exploration:

some basic Chart types

Bar chart

For categorical comparisons

 

Domain: categorical

Range: quantitative

data exploration:

some basic Chart types

Histogram

For categorical distributions

 

Domain: quantitative bins

Range: frequency of quantitative bin

data exploration:

some basic Chart types

Scatterplot

For correlation

 

2 categories, and the relationship between their quantitative values

data exploration:

some basic Chart types

Line chart

For temporal trends

 

Domain: temporal

Range: quantitative

data exploration:

vega-lite

A grammar of

interactive graphics

  • Data source
  • Mark (tick, bar, point, line, etc.)
  • Encoding (x, y, color, etc.)
  • For each encoding: type, field

data exploration

Brainstorm some charts

to answer the questions.

data exploration:

exercise (together)

Starter notebook

Full notebook

data exploration:
advice

  • Check for missing data, and the validity of the data
  • Focus on one question at a time (it's very easy to get sidetracked with a tangent)
  • If there IS an interesting tangent, make a note for later
  • If the question leads to a dead-end, explore another question or the tangent you found earlier
  • Don't be afraid to go out and look for additional data to aid your exploration
  • Sometimes, no interesting pattern IS very interesting

Translate from
data to design

  1. Concentrate on the takeaways to communicate across
  2. What does that mean in terms of the data?  (Individual or aggregate elements? Which attributes?)
  3. Map the relevant data to visual elements

Design:
marks & channels

Map individual or

aggregate data

elements to marks.

 

Map data attributes

to channels.

Design:
marks

Visualization Analysis and Design. Tamara Munzner, with illustrations by Eamonn Maguire. A K Peters Visualization Series, CRC Press, 2014.

Design:
channels

Visualization Analysis and Design. Tamara Munzner, with illustrations by
Eamonn Maguire. A K Peters Visualization Series, CRC Press, 2014.

Quantitative

  • Position
  • Size
  • Color

Categorical

  • Shape
  • Texture
  • Color

Temporal

  • Animation

Design:
marks & channels

Visualization Analysis and Design. Tamara Munzner, with illustrations by
Eamonn Maguire. A K Peters Visualization Series, CRC Press, 2014.

mark

bar

channels

x: category

y: quant

mark

point

channels

x: quant

y: quant

mark

point

channels

x: quant

y: quant

color: category

mark

point

channels

x: quant

y: quant

color: category

size: quant

Design:
channel effectiveness

Visualization Analysis and Design. Tamara Munzner, with illustrations by
Eamonn Maguire. A K Peters Visualization Series, CRC Press, 2014.

Design:
marks & channels

  • One-to-one mapping of data to channel

  • Multiple mappings of channel to mark (x, y, size, color usually)

  • Do not EVER map multiple data attributes to the same channel

Design:
Gestalt laws of grouping

the human mind naturally

groups individual elements

into patterns

use in data visualization to

save processing time

 

 

Design:
Gestalt laws of grouping

Proximity

Put related objects near each other

(The Functional Art, Ch. 6 by Alberto Cairo)

Design:
Gestalt laws of grouping

Similarity

Indicate like objects (helpful if they can't be placed close to each other)

(The Functional Art, Ch. 6 by Alberto Cairo)

Design:
Gestalt laws of grouping

Enclosure

Helpful when creating visualizations with multiple sections

(The Functional Art, Ch. 6 by Alberto Cairo)

Design:
remix & overlay

"You don’t always need to start from scratch, remix what’s out there already"

- Nadieh Bremer

But make sure they're the right visuals to communicate your message!

design:
exercise

Sketch all the things!

  1. What is your main message(s)?
  2. What marks will you use?  Do they represent individual data points, or aggregate?
  3. What channels will your marks use?  How do they support your message?

code:
design to code

  • Break it down!  What do you need to draw the marks?  What do you need to calculate the channels?

  • To draw marks: SVG (or canvas)

  • To calculate channels: D3 scales, shapes, and layouts (or straight-up math!)

SVG Elements

rect
x: x-coordinate of top-left
y: y-coordinate of top-left
width
height

circle
cx: x-coordinate of center
cy: y-coordinate of center
r: radius

text
x: x-coordinate
y: y-coordinate
dx: x-coordinate offset
dy: y-coordinate offset
text-anchor: horizontal text alignment

Hi!

path
d: path to follow

Moveto, Lineto, Curveto, Arcto

code:
D3 api

For translating raw data to what SVG needs to draw

Take output of layout calculations and draw SVG elements

Sometimes all you need are scales to get from data to screen space

Often times, you may need specific layouts.

These output x/y positions

And these generate path commands

Great dataviz-specific interactions

Code:
exercise (TOGETHER)

Implement the designs!

(Starter code)

Each curve represents a movie

  • x: release date

  • y: box office relative to mean

readability

Titles, descriptions, and legends

to explain the visualization

 

Axes and annotations

to describe the data

READABILITY:
axes & legends

d3-legend by Susie Lu

READABILITY:
Annotations

d3-annotation by Susie Lu

READABILITY:
exercise (TOGETHER)

Implement the designs!

(Starter code)

  1. Add axes
  2. Add legends
  3. Add annotations

more svg for

context & aesthetics

  • Patterns
  • Gradients
  • Text on a path
  • SVG filters
    (blurs, drop-shadows)
  • Clipping & masking

more svg:
exercise (TOGETHER)

Implement the designs!

(Starter code)

  1. Add textures for holidays
  2. Add dropshadow to movies

interactions

D3:

hover, click, and

other simple interactions

 

D3 & React (or similar):

  • Update after user manipulation of underlying data
  • Link multiple visualizations
  • Exploratory tools (filtering, aggregating)

final

visualization

exploratory

Process (at Netflix with Susie & Elijah):

  1. Initial meeting with stakeholders to figure out most important questions
  2. Meeting with data engineers
  3. Mock-up in sketch, sandbox with Semiotics, see shape of data
  4. Prototype, iterate with stakeholders

exploratory

Advice (from Netflix):

  • Different data sources, often SQL queries → plan for queries that take longer (important for interactions)
  • Questions for stakeholders:
    • What's the business question they're trying to answer?
    • How do the metrics they're comparing fall into a decision?

exploratory

Advice, cont.:

  • Level of aggregation that's most effective for decision making:
    • Get it to ~7 things that are granular and meaningful enough
    • If not, top 10 of a default
  • Gain trust and credibility within org
    • Have to compete with tables of data (detailed but hard to read)
    • People will get to a state you never designed for, so think through edge cases

resources

Books:

The Functional Art by Alberto Cairo

Visual Analysis and Design by Tamara Munzner

 

Online:

Datawrapper Blog

Flowing Data

The little of visualization design

The Pudding

Information is Beautiful Awards

 

Thank you's

Lisa Rost

Nadieh Bremer

Susie Lu

Elijah Meeks

 

Beta-testers

Kristin Henry

Micah Stubbs

Radames Ajna

Santhosh Soundararajan

Nathan Harris

Dylan Wootton

Made with Slides.com