Visual Analytics Lecture BigData Analytics Julian M. Kunkel julian.kunkel@googlemail.com University of Hamburg / German Climate Computing Center (DKRZ) 2017-10-20 Disclaimer: Big Data software is constantly updated, code samples may be outdated.
Visual Data Analysis Visual Perception Designing Graphics Summary Outline 1 Visual Data Analysis 2 Visual Perception 3 Designing Graphics 4 Summary Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 2 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Statistical Graphics [44] Definition: Graphics in the field of statistics used to visualize quantitative data Objectives The exploration of the content of a data set The use to find structure in data Checking assumptions in statistical models Communicate the results of an analysis Plots (Excerpt) Scatter, box, histograms Statistical maps Probability plots Source: William Playfair’s Time Series of Spaghetti plots Exports and Imports of Denmark and Norway [44] Residual plots Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 3 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Visual Analytics [32] Definition [33] The science of analytical reasoning facilitated by interactive visual interfaces. Objective Solve complex questions/time critical problems applying the scientific method Present gained insight / communicate it visually Analytical tasks Understanding past situations; trends and events that caused current conditions Monitoring events for indicators for an emergency Identifying possible alternative future scenarios and their warning signs Determining indicators of the intent of an action or an individual Supporting decision makers in times of crisis Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 4 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Visual Analytics Workflow Data Exploration User Interaction Visualization Perception Transformation Cognition Mapping Data Knowledge Data Mining Building Visualization Cognition Models Parameter refinement Automated Data Analysis Feedback loop Figure based on [48] Motto: Analyse First – Show the Important; Zoom, Filter and Analyse Further – Details on Julian M. KunkelDemand[34] Lecture BigData Analytics, WiSe 17/18 5 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Fields of Visual Analytics Source: Visual analytics: Scope and challenges [48] Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 6 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Human-Computer Interaction Why do we team humans and computers using a visual interface? Comparing capabilities of humans and computers Human brain processing power is enormous 100 billion neurons, linked together by many synapses Synapses fire with 4.3 · 1015 spikes/s; data rate of 1.1 · 1016 bits/s = 125 TiB/s; 20 Watt [6] Fastest supercomputer in the world [7]: Sunway TaihuLight: 125 TFlop/s, 15 MW Estimation: Simulating one second of human brain activity requires 83k processors Strength of humans and computers: Human Computer Pattern recognition Execution of algorithms Creative thinking Accuracy Processing new infos Visual perception and analysis capabilities exceed computers, e.g., computer vision Vision uses 30-50% of the brain’s capabilities ⇒ Visual representation and analytics is key for efficiency Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 7 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Example Analysis Session: Demo Based on a real case [35] 1854, Broad Street, London Within a few days people died mysteriously Dr. John Snow investigated the cause to stop “disease” He analyzed data visually with the scientific method We will follow his analysis steps Using modern data analytics tools Interactive lab notebook Record notes/hypothesis, type code, store it together with results The notebook is prepared using Jupyter with Python Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 8 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Analysis Results John found the source of the Cholera: The pump He claimed the disease is spread by the water John is one of the founders of our Germ theory They unmounted the pump handle But could not proof theory Board of health did not believe his analysis They believed “Miasma” is the cause ⇒ Convincing documentation is important! Original map made by John Snow in 1854. Cholera cases are highlighted in black. [35] Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 9 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary 1 Visual Data Analysis 2 Visual Perception 3 Designing Graphics 4 Summary Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 10 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Cognition Definition: The mental action or process of acquiring knowledge and understanding through thought, experience, and the senses [46] Communicated information and interpretation is biased by humans due to: Perception Information processing Subjective knowledge Psychology knows many cognitive biases [40] Categories of cognitive biases: Limits of memory Too much information Not enough meaning Need to act fast Categories serve as guidelines for visual analytics Source: Wikipedia’s complete (as of 2016) list of cognitive biases, beautifully arranged and designed We will focus on visual perception by John Manoogian III (jm3). Categories and descriptions originally by Buster Benson. [40] Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 11 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Visual Perception: Information Pipeline Information Communication Information is transformed several times from digital data to human The retina and brain interprets visual information Efficient communication requires to understand human perception Computer Human x, y,262,250, 1854-08-21, merchant, female298,242, 1854-08-24, other, child295,229, 1854-08-30, craftsmen, child270,249, 1854-08-31, service, man273,263, 1854-08-31, craftsmen, childdeathDate, job,Datagender > Display > Retina > Brain 234,236, 1854-09-01, service, child 252,253, 1854-09-01, service, child 305,221, 1854-09-01, scholar, man 275,238, 1854-09-01, service, elderly 251,245, 1854-09-01, scholar, child 282,259, 1854-09-01, other, child 263,247, 1854-09-01, military, elderly 237,239, 1854-09-01, service, elderly Electric 273,232, 1854-09-01, other, child 290,228, 1854-09-01, scholar, female Electric Light Chemic Visualization Perception Interpretation Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 12 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Optical Illusions [38] Sphere of the eye Retina Optic nerve Definition: visually perceived images that Crossed fibers Uncrossed fibers differ from objective reality Optic chiasma Optic tract Commissure of Gudden They are caused by the visual system They are many different types of illusions Pulvinar Lateral geniculate body Perceived colors and contrasts Superior colliculus Medial geniculate body Size and shapes of objects Nucleus of oculomotor nerve Interpretation of objects Nucleus of trochlar nerve Depth perception Nucleus of abducent nerve Moving of objects Afterimages ... Cortex of occipital lobes Source: Gray’s Anatomy depiction of the optic nerves & nuclei... KDS444 [39] Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 13 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Color Illusion Field A and B have the same gray tone Source: The checker shadow illusion. Edward H. Adelson Proof: Breaking the illusion. [38] Source: Edward Adelson [38] Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 14 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Color Illusion (2) Form that seems to be filled in yellow instead of white Source: Blue-bordered cookie that misleadingly seems to be filled with light yellow water-color. Jochen Burghardt. [38] Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 15 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Shapes of Objects Both orange circles are the same size Source: Optical illusion: The two orange circles are the same size. [38] Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 16 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Shapes of Objects (2) Vertical and horizontal lines have the same length Source: Vertical–horizontal illusion, S-kay [38] Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 17 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Shapes of Objects (3) Imaging a white triangle in the center Source: Kanizsa triangle. Fibonacci [38] Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 18 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Interpretation of Images Vase or two faces Source: Two silhouette profiles or a white vase?, Brocken Inaglory [38] Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 19 / 28
Visual Data Analysis Visual Perception Designing Graphics Summary Interpretation of Images (2) Duck or rabbit Source: Jastrow, J. (1899). The mind’s eye. Popular Science Monthly, 54 Julian M. Kunkel Lecture BigData Analytics, WiSe 17/18 20 / 28
Hiển thị 20 trang đầu. Tải tài liệu để xem đầy đủ.
Chưa có bình luận nào. Hãy là người đầu tiên!
Bình luận (0)
Chưa có bình luận nào. Hãy là người đầu tiên!