Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Problem Solving with Data Visualization

Problem Solving with Data Visualization

If we only have 60 minutes left, we are going to spend 55 minutes to think about the problem, while the last 5 minutes for the data visualization.

Fiqry Revadiansyah

August 20, 2021
Tweet

More Decks by Fiqry Revadiansyah

Other Decks in Design

Transcript

  1. Data Visualization
    Problem Solving with
    Presented by Fiqry R

    View Slide

  2. Hey I’m,
    Fiqry Revadiansyah
    I am a,
    ● CURRENTLY | Data Science @Accenture
    ● Previously | RnD Data Science @Purwadhika
    ● Firstly | Data Science @Bukalapak

    View Slide

  3. 01
    The Stupendous
    Problem Solving
    Act as a human, treat
    problem as a human
    02
    The Amalgamation of
    Cognitive and Affective
    Uninterruptible of
    combination arts
    03
    Simplicity is The
    Ultimate Sophistication
    Simpler, better
    Today’s Journey Path

    View Slide

  4. “If I had an hour to solve a
    problem, I would spend 55
    minutes thinking about
    the problem, and 5
    minutes thinking about
    the solutions”- You know him

    View Slide

  5. The Stupendous
    Problem Solving

    View Slide

  6. Think as Human

    View Slide

  7. Human Decisions
    are Emotional
    Kahneman, Daniel. 2011. Thinking Fast and Slow. New York: Farrar, Straus and Giroux (p.280)

    View Slide

  8. Act as Human
    Lerner, J., Li, Y., Valdesolo, P., & Kassam, K. (2015). Emotion and decision making. Annual review of psychology, 66, 799-823 .

    View Slide

  9. Emotions are Important
    in Decision Making
    Emotions help optimally
    navigate social decisions *)
    Emotions which are essential
    to build an impactful decision
    making: Memorable,
    Persuasive, and Engaging **)
    01 02
    *) Lerner, J., Li, Y., Valdesolo, P., & Kassam, K. (2015). Emotion and decision making. Annual review of psychology, 66, 799-823
    **) https://www.forbes.com/sites/brentdykes/2016/03/31/data-storytelling-the-essential-data-science-skill-everyone-needs/?sh=628b20e852ad

    View Slide

  10. Decision Theory Application
    In the reality, we already done this, but...
    Problem(s)
    Solution A
    Solution B
    Solution C
    Action A
    Action B
    Action C
    Result A
    Result B
    Result C
    Outcome(s)
    Incidental
    Influence(s) Emotion(s) Decision(s) Result(s)
    Possible
    Outcome(s)
    Business Problem Research & Discussion Data Visualization Data Insight Storytelling Leader’s Action
    [Untouchable Area]
    [Touchable Area]

    View Slide

  11. 01
    Different
    Outcome(s)
    Expectation mismatch
    to the reality
    02
    Disconnectedness
    Solution(s)
    Irrelevant approach to
    solve the problem
    03
    Graveyard
    Decision(s)
    Leader distrust our
    approach due to some
    mistakes in our insights
    Sometimes, we are
    experiencing ...

    View Slide

  12. Human-centered Design Framework (I)
    Put “Emotion” to all segments
    Problem(s)
    Solution A
    Solution B
    Solution C
    Action A
    Action B
    Action C
    Result A
    Result B
    Result C
    Outcome(s)
    Emotion(s) +
    Incidental
    Influence(s)
    Emotion(s)
    Emotion(s) +
    Decision(s)
    Emotion(s) +
    Result(s)
    Business Problem Research & Discussion Data Visualization Data Insight Storytelling
    Possible
    Outcome(s)
    Leader’s Action
    [Untouchable Area]
    [Touchable Area]

    View Slide

  13. Human-centered Design Framework (II)
    Stupendous Problem Solving
    Emotional
    Business Problem
    (Empathy)
    Emotional
    Discussion/
    Research
    Emotional
    Data Visualization
    Emotional
    Data Insight Storytelling
    * ) Ideo.org/tools

    View Slide

  14. HCD Frameworks
    will aim you to:
    ❏ Know where to begin with
    ❏ Remind human as a social creature
    ❏ Follow a continuously improve
    ❏ Create a high quality decision

    View Slide

  15. Human-centered Design Framework (III)
    What will we cover today
    Emotional
    Business Problem
    (Empathy)
    Emotional
    Discussion/
    Research
    Emotional
    Data Visualization
    Emotional
    Data Insight Storytelling
    PART II PART III
    How to pour “Emotion”
    to understand business problem
    How to pour “Emotion”
    to deliver insight and visualization

    View Slide

  16. “Design is not what it looks
    and feels like.
    Design is how it works”
    - You know him

    View Slide

  17. The Amalgamation
    of Cognitive &
    Affective

    View Slide

  18. Value vs
    Perceived Value
    Common Pitfalls of
    Understanding the Business Problem
    ❏ Individual Belief != Collective Belief
    ❏ “I think I am right” mindset
    ❏ Boiling the Ocean

    View Slide

  19. What we think is the best result
    for them (Cognitive)
    What they think is the best result
    for them (Cognitive)
    Value Perceived Value
    Empathy as a Bridge: Affective
    “I think, in order to increase revenue
    we should increase the sales”
    “We think, tackling fraud is the best way
    to reduce cost and gain more revenue”
    We need to pour our empathy to bridge communication, listen what they
    need, understand the situation they face, know the key priority, and so on.

    View Slide

  20. A • mal • ga • ma • tion
    /əˌmalɡəˈmāSH(ə)n/
    (Noun): The action, process, or result of combining or uniting
    Combining both Cognitive (Rational Thinking) and Affective (Emotional Thinking)
    will produce high-quality problem solving skill

    View Slide

  21. Amalgamation helps
    optimally navigate
    decision
    Emotions help optimally
    navigate social decisions *)
    Emotions which are essential
    to build an impactful decision
    making: Memorable,
    Persuasive, and Engaging **)
    01 02
    *) Lerner, J., Li, Y., Valdesolo, P., & Kassam, K. (2015). Emotion and decision making. Annual review of psychology, 66, 799-823
    **) https://www.forbes.com/sites/brentdykes/2016/03/31/data-storytelling-the-essential-data-science-skill-everyone-needs/?sh=628b20e852ad

    View Slide

  22. The First Principle Thinking (I)
    A Framework about how not to boil the ocean
    Analogy
    Electric cars will never work because
    batteries are expensive
    First Principle
    Purchasing individual battery materials
    from cost-efficient sources allows us to
    make batteries that are much cheaper
    than anyone realizes
    Grass, Kacper. 2019. First Principles Thinking for Societal Problem-Solving: A Manual to Generate Innovative Solutions to Today’s Challenges. Factory for Innovative Policy Solutions.
    https://doi.org/10.13140/RG.2.2.16182.24642/1

    View Slide

  23. “I tend to approach things
    from a physics framework,
    and physics teaches you to
    reason from first principles
    rather by analogy”
    - You know him

    View Slide

  24. The First Principle Thinking (II)
    Issue Tree: How to decompose problem into shattered pieces
    How to
    Increase
    Revenue
    Increase
    The Sales
    Optimize
    the Cost
    Increase
    Buyer
    Increase
    Variety
    Reduce
    Fraud
    Optimize
    Operational
    * ) Minto, Barbara. 1991. The Pyramid Principle, Logic in Writing and Thinking. Essex: Pearson Education Limited (p. 96)
    High Effort, Not Priority
    Possible, Not Priority
    Possible, Priority
    Pour Empathy by Listen to Their Situation

    View Slide

  25. The First Principle Thinking (III)
    Pitfall on Actionless Insight: Incomplete Point of View
    How to
    Increase
    Revenue
    Increase
    The Sales
    Optimize
    the Cost
    Increase
    Buyer
    Increase
    Variety
    Reduce
    Fraud
    Optimize
    Operational
    “Insights on Fraudsters Data”
    A question arise.
    Is this enough?

    View Slide

  26. The Machine Learning Principle (I)
    Interconnected data: Opens a new horizon of understanding

    View Slide

  27. The Machine Learning Principle (II)
    Train a model to distinguish whether a photo is about Cat or Dog
    Cat Picture Dog Picture

    View Slide

  28. The Machine Learning Principle (III)
    Train a model to distinguish whether a photo is about Cat or Dog
    Cat Picture
    Dog Picture
    Similarity Dissimilarity
    4 Legs Skin pattern
    2 eyes Food
    Has a tail Behavior
    Has a neck Work in Police Station
    Has no wings Family/Ordo
    Trial 1 Training Feature
    5 features: 70% Accuracy
    Trial 2 Training Feature
    10 features: 88% Accuracy
    Trial 3 Training Validated Feature
    8 features: 94% Accuracy
    Training More Validated Features
    Higher Model Accuracy

    View Slide

  29. The Machine Learning Principle (IV)
    Add more features to increase the “accuracy” of the insights
    Additional Insights
    “New user referrals who made transaction
    on the same day of their registration date are
    suspected as fraudsters by XX%” Sunburst Plot

    View Slide

  30. Simplicity is the
    Ultimate Sophistication

    View Slide

  31. Breathtaking visualization
    helps to build impactful
    decision
    Emotions help optimally
    navigate social decisions *)
    Emotions which are essential
    to build an impactful decision
    making: Memorable,
    Persuasive, and Engaging **)
    01 02
    *) Lerner, J., Li, Y., Valdesolo, P., & Kassam, K. (2015). Emotion and decision making. Annual review of psychology, 66, 799-823
    **) https://www.forbes.com/sites/brentdykes/2016/03/31/data-storytelling-the-essential-data-science-skill-everyone-needs/?sh=628b20e852ad

    View Slide

  32. 01
    Data Ink Ratio
    02
    Gestalt Principle
    03
    Plot Choice
    Breathtaking Visualization
    Elements

    View Slide

  33. Data Ink
    Ratio
    Simplicity through Data Ink
    The non-erasable core of a graphic
    * ) Tufte, Edward R. 2001. The Visual Display of Quantitative Information. Connecticut:
    Graphics Press LLC
    “Above all else show the data”
    - Edward R. Tufte (p.92)

    View Slide

  34. Data Ink Ratio (I)
    Confession of Sin: Tell us all of the sins from this graph?

    View Slide

  35. Data Ink Ratio (II)
    What can be improved?

    View Slide

  36. Data Ink Ratio (III)
    What else can be improved? (Remove Background)

    View Slide

  37. Data Ink Ratio (IV)
    What else can be improved? (Remove border)

    View Slide

  38. “The whole is other than the sum of the parts.”
    -Kurt Koffka
    Gestalt Principle
    Simplicity through Human Psychology. Gestalt is a psychology term
    which means “unified whole”, refers to theories of visual perception
    developed by German psychologist in the 1920s.
    http://www.gestalttheory.net/cms/uploads/pdf/archive/1910_1933/gestalt_theory_wertheimer.pdf

    View Slide

  39. Gestalt Principles
    Elements of Gestalt Law

    View Slide

  40. Gestalt Principles (I)
    Law of Continuation: Elements arranged in a line or soft curve are perceived
    to be more related than those arranged randomly or in a harsh line

    View Slide

  41. Gestalt Principles (I)
    Law of Continuation: Elements arranged in a line or soft curve are perceived
    to be more related than those arranged randomly or in a harsh line

    View Slide

  42. Gestalt Principles (I)
    Law of Continuation: Elements arranged in a line or soft curve are perceived
    to be more related than those arranged randomly or in a harsh line

    View Slide

  43. Gestalt Principles (II)
    Law of Proximity: Elements arranged close to each other are perceived
    as more related than those placed further apart

    View Slide

  44. Gestalt Principles (II)
    Law of Proximity: Elements arranged close to each other are perceived
    as more related than those placed further apart

    View Slide

  45. Gestalt Principles (II)
    Law of Proximity: Elements arranged close to each other are perceived
    as more related than those placed further apart
    Company A
    Company B

    View Slide

  46. Gestalt Principles (III)
    Law of Similarity: Elements sharing similar visual characteristics are
    perceived to be more related than those not sharing similar characteristics

    View Slide

  47. Gestalt Principles (III)
    Law of Similarity: Elements sharing similar visual characteristics are
    perceived to be more related than those not sharing similar characteristics
    Company A
    Company B

    View Slide

  48. Gestalt Principles (III)
    Law of Similarity: Elements sharing similar visual characteristics are
    perceived to be more related than those not sharing similar characteristics
    Company A
    Company B

    View Slide

  49. Gestalt Principles (IV)
    Law of Common Fate: Elements placed within the same region or same
    background are perceived as grouped

    View Slide

  50. Gestalt Principles (IV)
    Law of Common Fate: Elements placed within the same region or same
    background are perceived as grouped
    Company A
    Company B

    View Slide

  51. Gestalt Principles (IV)
    Law of Common Fate: Elements placed within the same region or same
    background are perceived as grouped
    Company A
    Company B

    View Slide

  52. Final Touch for Data Visualization
    Even though company A started to sell four times more than B on January, at the end
    their performance outnumbered by three times just within a year.
    Company A
    Company B
    40
    125
    180
    60
    158
    Leadership Change of Company B: Sales are Skyrocketing

    View Slide

  53. Plot Choice
    Simplify by Plot Function. Type of Data Visualization
    https://www.mindbowser.com/data-visualization-how-to-use-it-to-your-advantage/
    and some google image pictures
    Use Sankey/Sunburst/Treemap to show journey of grouped-data

    View Slide

  54. Plot Choice
    Targeted plot choice will align with the
    needs of business problem

    View Slide

  55. “Simplicity is the
    Ultimate Sophistication”
    - Not all of you know him

    View Slide

  56. To maximize the outcome from your
    problem, you should have to treat
    them as a human: Pour Emotion
    Summary
    - The First Principle Thinking
    - Feature Engineering Principle
    - Data Ink Ratio
    - Gestalt Principle

    View Slide

  57. THANK YOU

    View Slide

  58. Credits
    Books and Academic Materials
    ● Grass, Kacper. 2019. First Principles Thinking for Societal Problem-Solving: A Manual to Generate Innovative Solutions to
    Today’s Challenges. Factory for Innovative Policy Solutions. https://doi.org/10.13140/RG.2.2.16182.24642/1
    ● Kahneman, Daniel. 2011. Thinking Fast and Slow. New York: Farrar, Straus and Giroux (p.280)
    ● Lerner, J., Li, Y., Valdesolo, P., & Kassam, K. (2015). Emotion and decision making. Annual review of psychology, 66, 799-823
    ● Minto, Barbara. 1991. The Pyramid Principle, Logic in Writing and Thinking. Essex: Pearson Education Limited (p. 96)
    ● Tufte, Edward R. 2001. The Visual Display of Quantitative Information. Connecticut: Graphics Press LLC
    Articles and Websites
    ● ideo.org/tools
    ● https://www.forbes.com/sites/brentdykes/2016/03/31/data-storytelling-the-essential-data-science-skill-everyone-needs/?sh=62
    8b20e852ad
    ● http://www.gestalttheory.net/cms/uploads/pdf/archive/1910_1933/gestalt_theory_wertheimer.pdf
    ● https://medium.muz.li/gestalt-principles-in-ui-design-6b75a41e9965
    ● https://www.mindbowser.com/data-visualization-how-to-use-it-to-your-advantage/
    ● https://towardsdatascience.com/little-known-ways-to-make-your-data-visualization-awesome-890d71b6e365
    ● https://s3.cad.rit.edu/cadgallery_production/storage/media/uploads/faculty-s-projects/2130/documents/243/gestalt-principles-
    keli-dirisio.pdf
    ● http://www2.jufejus.org.ar/www.jufejus.org.ar/images/doc/ACTIVIDADES/Estadisticas/Jornadas%20de%20Capacitacion/Materi
    al/JornadasX/Tufte/The%20Visual%20Display%20of%20Q%20Info.pdf

    View Slide