Why: abstracting the tasks
Information Visualization

What we are going to learn

  • Abstracting the tasks (why)
  • Define the verb of the task (action)
  • Define the noun of the task (targets)
  • Tasks good practices
  • Deriving new data

Abstracting the tasks (Why)

Why analyze tasks?

  • Abstract tasks are domain-independent
  • Two different domain problems πŸ‘‰ same task abstraction πŸ‘‰ same solution
  • Visualization idioms are good for some tasks and bad for others

Task construction

Action (verb)
Why Actions
+
Target (noun)
Why Targets

Actions
verbs

Actions: Levels

Three levels:
  • Analyze
  • Search
  • Query
Try being specific

Actions: Analyze, High-level

  • Consume
    • Discover vs. present
    • Enjoy
  • Produce
    • Annotate, record
    • Derive
Why Actions: Analyze

Actions: Search, Mid-level

Why Actions: search
  • Target
    • I know what I'm looking for
    • E.g., what's the salary of the CEO? πŸ‘† Target known
    • E.g., who has the highest salary? πŸ‘† Target unknown

Actions: Search, Mid-level

Why Actions: search
  • Location
    • I know where the mark is going to be located in the chart
    • Scales ordered alphabetically

Actions: Query, Low-level

How much of the data matter?
  • One πŸ‘‰ identify
  • Some πŸ‘‰ compare
  • All πŸ‘‰ summarize
Why Actions: query

Targets
nouns

Why Targets

Targets: Trends, outliers, features

  • Trends (or patterns): e.g., are sales increasing?
  • Outliers: one or some that aren't like the others
  • Features: πŸ‘‰ task-dependent (e.g., a V-shaped recession)
Why Targets: All data

One Attribute: Distributions, Extremes

Why Targets: Attributes
  • Distribution (e.g., GDP distribution across multiple countries)
  • Extremes (e.g., country with highest GDP)

Many Attributes: Dependency, Correlation, Similarity

  • Dependency (e.g., do all NBA players attended college?)
  • Correlation (e.g., do higher-paid players score more?)
  • Similarity (e.g., do Curry and Hardem shoot the same?)
Why Targets: Attributes

Networks: Topology and Shape

  • Topology (e.g., do all my classmates follow each other?)
  • Paths (e.g., who do I know on LinkedIn that knows Bill Gates?)
Why Targets: Networks

Targets: Spatial data

Why Targets: Networks
  • Shape (e.g., is there a park within 100 feet of every school?)

Good practices

  • Be specific
  • Is your task measurable?
  • Number your tasks
  • Use Tamara's grammar
  • Describe specific attributes used
  • Don't think about the visuals
  • T1: Summarize the stock market prices for the S&P 500 companies (Summarize, Distribution)
  • T2: Identify the most profitable books for the last year (Identify, Features)
  • T3: Compare the trends of COVID-19 deaths in the world (Compare, Trend)

Abstraction examples:

Policy makers in California want to test what counties had flattened the curve in the Covid 19 pandemic.

T1. Compare the (shape of the curves of the) number of positive cases per day over time between counties (compare, trend/features)
Economists want to classify what type of recession was generated by the Covid 19 pandemic in different countries

T2. Compare the shape of the GDP over time per country (compare, trend/features)

Example insights

Santa Guerra County has been the best at flattening the curve in California, with a reduction of 37% of their expected deaths by June 15, 2020.

T1. Compare the (shape of the curves of the) number of positive cases per day over time between counties (compare, trend/features)

Example insights

  • Imagine what the visualization users could be able to discover
  • Made up
  • Include details (names and/or numbers)
  • Are very specific
  • Can be tested
  • Can be led to action

Deriving new data

Why: Derive data
Task: Summarize the trend of the trade balance (exports - imports)Derive data before
Don't make the user compute it in their head.
Task: Summarize the trend of the trade balance (exports - imports)Derive data after
Instead, display the difference.

When to derive data

  • Large data? πŸ‘‰ Aggregate
  • Complex data? πŸ‘‰ Simplify
  • Dimensionality reduction
  • Network/tree analytics (clusters, centralities)
  • Don't make the user think!
Why: Derive data

What we learned

  • Abstracting the tasks (why)
  • Define the verb of the task (action)
  • Define the noun of the task (targets)
  • Tasks good practices
  • Deriving new data

Practice

Rendering

SVG

https://youtu.be/6_pYq2o2P0c

Canvas

https://youtu.be/MvB-n9JVW3M

WebGL?

https://youtu.be/EGgRN7RjQrs

Setting up D3

https://youtu.be/2l7zpJUIBgo

Setting up Vega-lite

https://youtu.be/iFJRKfpof_w

Setting up Tableau

https://youtu.be/iFJRKfpof_w