Summary Semiotics Perception Data Jrg Cassens Representation - PowerPoint PPT Presentation

Types of Analysis Introduction � Statistical Analysis/Profiling History What are the entities that are being described (e.g. persons, Semiotics Communication grants, publications)? Semiotics Classification � Temporal Analysis: When Framework Perception Does the visualization show a development over time? Data � Geospatial Analysis: Where Representation Does the visualization include information about location? Presentation ≡ Topical Analysis: What References What is the topical area of the visualization? ▽ Network Analysis: With Whom Does the visualization contain information about social networks? SoSe 2017 Jörg Cassens – Summary 25 / 180

Audience Introduction History � Gender – are we targeting a certain gender? Semiotics ⑤ Age – is it intended for certain age groups? Communication Semiotics Classification � Education – is the level of education important Framework Perception � Disability – are disabilities taken into account (for example Data colour blindness)? Representation � Contextual parameters, e.g. Presentation � Leisure – related to our leisure References � Business – related to business � Scientific – related to science � Religious – related to religion � Any other information defining the audience SoSe 2017 Jörg Cassens – Summary 26 / 180

Medium Introduction History Semiotics ✎ Printed medium Communication Semiotics � Digital medium Classification Framework ✇ Time-based – visualizing information using time Perception ⊙ Location-based – spatially visualizing information Data Representation � Modality Text – contains text Presentation ֠ Modality Sound – contains sound References � Interactive visualization � Other – other information about the medium SoSe 2017 Jörg Cassens – Summary 27 / 180

Framework Introduction Level Audience Medium History � Micro level � Gender ✎ Printed Semiotics Communication � Meso level ⑤ Age � Digital � Semiotics Classification Framework � Macro level � Education ✇ Time-based Perception � Disability ⊙ Spatial Data Type � Context, e.g. � With Text Representation Presentation � Profiling � Leisure ֠ With Sound � Business References � Temporal � Interactive � Scientific � Geospatial � Religious � Other � Other ≡ Topical ▽ Network SoSe 2017 Jörg Cassens – Summary 28 / 180

Area Introduction History Semiotics Perception Physiology Color Processing Pipeline Attention PERCEPTION Data Representation Presentation References SoSe 2017 Jörg Cassens – Summary 29 / 180

Topic Introduction History Semiotics Perception Physiology Color Processing Pipeline Attention Physiology Data Representation Presentation References SoSe 2017 Jörg Cassens – Summary 30 / 180

Gesichtsfeld Introduction History Semiotics Perception Physiology Color Processing Pipeline Attention Data Representation Presentation Source: Malaka et al. (2009) References Höchste Auflösung in der Fovea in der Mitte des Sehfeldes Dort finden sich viele Zapfen, aber keine Stäbchen Nachts sind wir im Zentrum des Sehfeldes faktisch blind In der Peripherie ist das Sehen stark eingeschränkt SoSe 2017 Jörg Cassens – Summary 31 / 180

Retina Introduction History Semiotics Perception Physiology Color Processing Pipeline Attention Data Representation Presentation References Source: Zimbardo et al. (2012) SoSe 2017 Jörg Cassens – Summary 32 / 180

Rezeptives Feld Bei Ganglionzellen ist das rezeptive Feld rund Introduction Das rezeptive Feld wird in ein Zentrum und ein Umfeld History unterteilt und man unterscheidet On-Zentrum-Neurone Semiotics und Off-Zentrum-Neurone Perception On-Zentrum-Neuronen haben ein erregendes Zentrum und Physiology Color ein hemmendes Umfeld Processing Pipeline Bei Off-Zentrum-Neuronen verhält es sich umgekehrt Attention Data Durch Erregung und Hemmung wird die Feuerrate des Representation Neurons manipuliert Presentation References Licht (weiß), On-Zentrum Kante, On-Zentrum Source: Jänicke (2016) Source: Ware (2004) SoSe 2017 Jörg Cassens – Summary 33 / 180

Optische Täuschungen Mit dieser Theorie kann man einige optische Täuschungen Introduction erklären History Hermann Gitter (links): Schwarze Punkte erscheinen an Semiotics Perception den Schnitten weißer Geraden Physiology Color Kontrast Illusion (rechts): Abhängig von der Processing Pipeline Hintergrundfarbe wird ein und derselbe Grauton Attention Data unterschiedlich wahrgenommen Representation Presentation References Source: Ware (2004) SoSe 2017 Jörg Cassens – Summary 34 / 180

Spatial Contrast Introduction History Semiotics Perception Physiology Color Processing Pipeline Attention Data Representation Presentation References Source: Ware (2004) SoSe 2017 Jörg Cassens – Summary 35 / 180

Topic Introduction History Semiotics Perception Physiology Color Processing Pipeline Attention Color Data Representation Presentation References SoSe 2017 Jörg Cassens – Summary 36 / 180

Isoluminanz Introduction History Semiotics Perception Physiology Color Processing Pipeline Attention Data Representation Presentation References Source: Ware (2004) SoSe 2017 Jörg Cassens – Summary 37 / 180

Isoluminanz Introduction History Semiotics Perception Physiology Color Processing Pipeline Attention Data Representation Presentation References Source: Ware (2004) SoSe 2017 Jörg Cassens – Summary 38 / 180

Farbkodierung von Objekten Introduction History Bei der Verwendung von Farbe zur Unterscheidung von Semiotics Merkmalen müssen einige Punkte beachtet werden: Perception Physiology Unterscheidbarkeit Color Processing Pipeline Eindeutige Farbtöne Attention Data Kontrast zum Hintergrund Representation Farbschwäche Presentation Anzahl References Größe der Farbfläche Konventionen SoSe 2017 Jörg Cassens – Summary 39 / 180

Farbkodierung von Objekten Unterscheidbarkeit : Die Farben sollen leicht voneinander Introduction zu unterscheiden sein History Wenn es darum geht ein Objekt einer bestimmten Farbe Semiotics schnell zu finden, sollte diese außerhalb der konvexen Perception Hülle der anderen Farben liegen Physiology Color Processing Pipeline Attention Data Representation Presentation References Source: Jänicke (2016); Ware (2004) SoSe 2017 Jörg Cassens – Summary 40 / 180

Farbkodierung von Objekten Eindeutige Farbtöne : Gegenfarben haben in den meisten Introduction Kulturen und Sprachen einen eigenen spezifischen Namen History und werden leicht erkannt Semiotics Perception Zu bevorzugen, wenn nur wenige Farben benötigt werden Physiology Wenn möglich nicht mehrere Farben aus der gleichen Color Farbfamilie verwenden Processing Pipeline Attention Gegenfarben: Blau-Gelb, Rot-Grün, Schwarz-Weiß Data Representation Presentation References Families of colors: (a) Pairs related by hue, family members differ in saturation. (b) Pairs related by hue, family members differ in saturation and lightness. (c) A family of warm hues and a family of cool hues. Source: Ware (2004) SoSe 2017 Jörg Cassens – Summary 41 / 180

Farbkodierung von Objekten Introduction History Kontrast zum Hintergrund : Es muss beachtet werden, dass Farben auf unterschiedlichem Hintergrund Semiotics unterschiedlich wirken können Perception Physiology Wechselwirkungen können durch eine einheitliche Kontur Color Processing Pipeline (z.B. schwarz oder weiß) verkleinert werden Attention Isoluminanz zwischen Objekt und Hintergrund ist zu Data vermeiden Representation Farbschwäche : Da es relativ viele Menschen mit Presentation References Farbschwäche gibt sollten Farbkodierung basierend auf rot-grün Kontrasten vermieden werden Anzahl : Nur 5 bis 10 Farben können schnell unterschieden werden SoSe 2017 Jörg Cassens – Summary 42 / 180

Farbkodierung von Objekten Introduction Größe der Farbfläche : Die Größe der farblich kodierten History Objekte sollte nicht zu klein sein, da sie sonst nicht Semiotics unterschieden werden können. Perception Allgemein gilt: Für kleine Farbflächen sollten stark Physiology Color gesättigte und stark unterschiedliche Farben verwendet Processing Pipeline Attention werden, für große Flächen eher Farben mit niedrigerer Data Sättigung und geringerem Abstand Representation Bei farbig hinterlegtem Text sollte eine helle Farbe gewählt Presentation werden References Konventionen : Einige Farben haben bestimmte Bedeutungen Rot = heiß oder Gefahr – Blau = kalt – Grün = Leben Man beachte: Andere Länder, andere Sitten! z.B. in China gilt rot = Leben oder Glück und grün = Tod SoSe 2017 Jörg Cassens – Summary 43 / 180

Topic Introduction History Semiotics Perception Physiology Color Processing Pipeline Attention Processing Pipeline Data Representation Presentation References SoSe 2017 Jörg Cassens – Summary 44 / 180

Verarbeitungsprozesse Verarbeitung visueller Introduction Information komplexer History Prozess Semiotics Man unterscheidet grob drei Perception Stufen der Verarbeitung: Physiology Color Sensorische Prozesse 1 Processing Pipeline Attention Parallele Erfassung Data grundlegender Representation Merkmale Presentation 2 Perzeptuelle Organisation References Mustererkennung 3 Aufgabenorientierte Verarbeitung Identifikation Wiedererkennen Source: Ware (2004), Gerrig and Zimbardo (2008, Graphik) SoSe 2017 Jörg Cassens – Summary 45 / 180

Stufe 1: Sensorische Prozesse Milliarden Neuronen erfassen gleichzeitig unterschiedliche Introduction Merkmale des visuellen Feldes, z.B. Helligkeit, Farbe und History Orientierung von Kanten Semiotics Diese initiale Verarbeitung ist zum größten Teil unabhängig Perception davon, worauf wir unsere Aufmerksamkeit richten Physiology Color Wichtige Merkmale: Processing Pipeline Attention Schnelle parallele Verarbeitung Data Extraktion fundamentaler Merkmale Representation Information wird nur kurz gespeichert Presentation Datenbasiertes bottom-up Modell der Verarbeitung References In der ersten Stufe kann sehr viel visuelle Information parallel verarbeitet werden Kann genutzt werden um die Aufmerksamkeit zu lenken; bestimmte Aspekte hervorzuheben So kann man den Betrachter dabei unterstützen wichtige Informationen schnell zu erkennen. SoSe 2017 Jörg Cassens – Summary 46 / 180

Stufe 2: Perzeptuelle Organisation Schätzungen der wahrscheinlichen Größe, Form, Introduction Bewegung, Entfernung und Ausrichtung eines Objekts History Schätzungen basieren auf mentalen Berechnungen, die Semiotics Vorwissen mit aktueller Evidenz aus den Sinnen sowie dem Perception Reiz in seinem Wahrnehmungskontext kombinieren Physiology Color Synthese einfacher sensorischer Merkmale wie Processing Pipeline Attention beispielsweise Farben, Kanten und Linien zu einem Perzept Data eines Objekts Representation Wichtige Merkmale: Presentation References Langsame serielle Verarbeitung Verwendung von Kurzzeit- und Langzeitgedächtnis Wechsel zwischen Merkmalsverarbeitung (bottom-up) und Aufmerksamkeit (top-down) Symbole erhalten komplexere Bedeutungen Verschiedene Verarbeitungspfade Objekterkennung – what-system Bewegungssteuerung – action-system, where-system SoSe 2017 Jörg Cassens – Summary 47 / 180

Stufe 3: Aufgabenorientierte Verarbeitung Weist den Perzepten Bedeutung zu Introduction Runde Objekte “werden” zu Fußbällen, Münzen, Uhren, History Orangen oder Monden; Menschen werden als weiblich oder Semiotics Perception männlich identifiziert, Freund/Feind, Verwandter/Star Physiology Die Aufmerksamkeit wird gezielt auf relevant Aspekte des Color Processing Pipeline visuellen Feldes gerichtet und wenige relevante Objekte Attention Data werden im Kurzzeitgedächtnis gespeichert Representation Wichtige Merkmale: Presentation Langsame serielle Verarbeitung References Verwendung von Kurzzeit- und Langzeitgedächtnis Top-down Verarbeitung Verarbeitung richtet sich nach der Fragestellung Verschiedene Objekte in einer Visualisierung sollten deutlich unterscheidbar sein, um diesen Prozess zu beschleunigen (vergleiche “Wo ist Walter?”) SoSe 2017 Jörg Cassens – Summary 48 / 180

Topic Introduction History Semiotics Perception Physiology Color Processing Pipeline Attention Attention Data Representation Presentation References SoSe 2017 Jörg Cassens – Summary 49 / 180

Präattentive Verarbeitung: Definition Introduction Definition History Präattentive Verarbeitung: Die Verarbeitung sensorischer Semiotics Informationen, die einer Aufmerksamkeitszuwendung auf Perception spezifische Objekte vorausgeht (Gerrig and Zimbardo, 2008). Physiology Color Processing Pipeline Attention Die erste Stufe der Verarbeitung visueller Information Data erfasst das gesamte visuelle Feld Representation Presentation Dieser Schritt wird präattentiv genannt, da in ihm References Informationen erfasst werden noch bevor Aufmerksamkeit (attention) darauf gerichtet wird Ob ein Reiz präattentiv ist wird experimentell bestimmt, indem man die Zeit misst, die Testpersonen brauchen um den Zielreiz in einer Menge von Distraktoren zu finden SoSe 2017 Jörg Cassens – Summary 50 / 180

Präattentive Wahrnehmung: Beispiele Form: Ausrichtung Introduction Größe History Krümmung Semiotics Länge & Breite Perception von Linien Physiology Color Anzahl Processing Pipeline Attention Annotationen Data Farbe: Representation Farbton, Presentation Intensität References Räumliche Position: Konkav, Konvex Einschluß Nicht Parallelität Nicht Verbindung Source: Ware (2004) SoSe 2017 Jörg Cassens – Summary 51 / 180

Two-Streams-Theory Introduction Nach der “Two-Streams History Theorie” wird die optische Semiotics Information danach in zwei Perception Physiology Systemen weiterverarbeitet, Color dem dorsalen “Wo-System” Processing Pipeline Attention und dem ventralen Data “Was-System” Source: Ware (2004) Representation Presentation Das dorsale System (grün) ist u.a. für die Wahrnehmung References von Bewegung, Tiefe, räumlicher Organisation und für die Planung von Handlungen (z.B. Greifen) verantwortlich Das ventrale System (lila) gleicht die aufgenommene Information mit vorhandenem Wissen ab und ordnet das Gesehene ein – es ist u.a. verantwortlich für Wahrnehmung von Objekten, Formen und Gesichtern SoSe 2017 Jörg Cassens – Summary 52 / 180

Scheinwerfertheorie Introduction Die Scheinwerfertheorie erklärt wie die Aufmerksamkeit History beim Betrachten einer Szene gesteuert wird Semiotics Grundannahme ist dass die Aufmerksamkeit wie ein Perception Scheinwerfer ist, der verschiedene Aspekte einer Szene Physiology Color beleuchten kann Processing Pipeline Attention Fällt die Aufmerksamkeit des Betrachters auf einen kleinen Data Teil der Szene, kann man dort genaue Details wahrnehmen Representation Presentation Die Verarbeitung erfolgt seriell, so dass der References Aufmerksamkeitsscheinwerfer von einem Punkt zum nächsten geleitet wird Der “Weg” der Aufmerksamkeit durch die Szene ist kontextabhängig, z.B. ausgehend von der Aufgabe, die gelöst werden soll SoSe 2017 Jörg Cassens – Summary 53 / 180

Introduction History Semiotics Perception Physiology Color Processing Pipeline Attention Data Representation Presentation References Source: Jänicke (2016) SoSe 2017 Jörg Cassens – Summary 54 / 180

Gestaltgesetze (-prinzipien) Gesetz der Nähe Introduction gruppiert Dinge zusammen, die räumlich oder zeitlich nah History sind Semiotics Gesetz der Ähnlichkeit/Gleichheit Perception Physiology gruppiert Bildteile, die nach Farbe, Form, Helligkeit, Größe, Color Orientierung ähnlich sind Processing Pipeline Attention Gesetz der guten Fortsetzung Data präferiert räumliche oder zeitliche Einfachheit Representation Presentation Gesetz der Geschlossenheit References neigt dazu, kleine Lücken aufzufüllen Gesetz des gemeinsamen Schicksals Objekte werden gerne als Gruppen wahrgenommen Weiterhin komplexere Prinzipien Gesetz der Symmetrie Unterscheidung von Vorder- und Hintergrund SoSe 2017 Jörg Cassens – Summary 55 / 180

Area Introduction History Semiotics Perception Data Communicate Process Aspects of Data DATA Data Preparation Focus Representation Presentation References SoSe 2017 Jörg Cassens – Summary 56 / 180

Topic Introduction History Semiotics Perception Data Communicate Process Communicate Aspects of Data Data Preparation Focus Representation Presentation References SoSe 2017 Jörg Cassens – Summary 57 / 180

Purpose Moving on: What is the purpose of your visualization? Introduction History Be clear about the motivation behind a project’s inception Semiotics Involves identifying who it is for and what needs you are Perception trying to fulfill Data Communicate What is the intention behind your project and how do you Process Aspects of Data define the visualization’s function and tone Data Preparation Focus Identify and assess the impact of the additional key factors Representation that will have an effect on your project Presentation Helps you surface all the restrictions, characteristics, and References requirements surrounding your project that will determine how you tackle it What is a purpose? reason for existing intended effect SoSe 2017 Jörg Cassens – Summary 58 / 180

Intent: The Visualization’s Function Introduction The intended function of a data visualization concerns the History functional experience you create between your design, the Semiotics data, and the reader/user Perception Data We can form three separate clusters or categories of Communicate function Process Aspects of Data Data Preparation While there is always a chance of slight overlap, there will Focus be a significant difference in your design choices Representation depending on whether the function of your visualization is Presentation to: References Convey an explanatory portrayal of data to a reader Provide an interface to data in order to facilitate visual exploration Use data as an exhibition of self-expression SoSe 2017 Jörg Cassens – Summary 59 / 180

Explain Explanatory data visualization is about conveying Introduction information to a reader in a way that is based around a History specific and focused narrative Semiotics Editorial approach to synthesize the requirements of your Perception target audience with the key insights and most important Data analytical dimensions you are wishing to convey Communicate Process Different approaches: Aspects of Data Data Preparation Information dashboard in a corporate setting (performance Focus figures with problems highlighted) Representation A graphic in a newspaper, explaining the complexity and Presentation severity of the problems around the economic crisis References An animated design to display patterns of population migration over time Physical or ambient visualization designed to draw attention to the sugar content of certain drinks The end result is typically a visual experience built around a carefully constructed narrative SoSe 2017 Jörg Cassens – Summary 60 / 180

Explore I Exploratory data visualization design is slightly different: Introduction we are seeking to facilitate the familiarization and History reasoning of data through a range of user-driven Semiotics experiences Perception Data In contrast to explanatory-based functions, exploratory Communicate data visualizations lack a specific, single narrative Process Aspects of Data Data Preparation They are more about visual analysis than just the visual Focus presentation of data Representation Presentation Exploratory solutions aim to create a tool, providing the References user with an interface to visually explore the data They can seek out personal discoveries, patterns, and relationships, thereby triggering and iterating curiosities Opens up the possibility for chance or serendipitous findings caused by forming different combinations of variable displays SoSe 2017 Jörg Cassens – Summary 61 / 180

Explore II Introduction History Semiotics The key feature that differentiates an exploratory piece Perception from an explanatory piece is the amount of work you have Data to do as a reader to discover insights Communicate Process For explanatory pieces, the designer should do the hard Aspects of Data Data Preparation work and create a clear portrayal of the interesting stories Focus and analysis from a dataset Representation An exploratory piece will be more about the readers doing Presentation the analysis themselves, putting the effort in to discover References things that strike them as being significant or interesting SoSe 2017 Jörg Cassens – Summary 62 / 180

Exhibit Designs that use data as the raw material, but where the Introduction intention is somewhat removed from a pure desire to History inform Semiotics Rather, the objective is closer to a form of exhibition or Perception self-expression through data representation Data Communicate This genre of work embodies the term “data art” Process Aspects of Data Characterized by a lack of structured narrative and absence Data Preparation Focus of any visual analysis capability Representation Instead, the motivation is much more about creating an Presentation artifact, an aesthetic representation or perhaps a References technical/technique demonstration In the following example, we see an example of “data art” that visualizes all the adjectives used in Cormac McCarthy’s book “The Road” Adjectives arranged radially in alphabetical order, each line represents a timeline of the book, beginning at perimeter SoSe 2017 Jörg Cassens – Summary 63 / 180

Science vs. Art “Science” Introduction Concerned with preserving the efficiency and accuracy of History judgments derived from a visualization Semiotics Variations in data representation that steer away from this Perception goal are believed to reduce the quality and effectiveness of Data a visualization Communicate Process “Art” Aspects of Data Data Preparation Concerned with experimentation, finding creative Focus Representation expressions of data, and new aesthetic connections with an Presentation audience References The latter enhances the field by demonstrating what can be achieved through the aesthetic and technological creativity The former help us understand what we should do through the pursuit of evidence and observation of rules around human cognition and visual perception SoSe 2017 Jörg Cassens – Summary 64 / 180

Pragmatic and Analytical Introduction Jock Mackinlay History Semiotics A visualization is more effective than another visualization if the Perception information conveyed by one visualization is more readily Data perceived than the information in the other (in: Kirk (2012)). Communicate Process Aspects of Data Data Preparation Designs that fit this classification will ofen involve data Focus being represented through the use of bar charts, line charts Representation Presentation and dot plots, for example References Stylistically, they will be characterized by a rather clinical look-and-feel that is consistent with the next sample image, taken from a project analyzing Olympic results over the years SoSe 2017 Jörg Cassens – Summary 65 / 180

Emotive and abstract Chris Jordan Introduction History I have a fear that we aren’t feeling enough, we aren’t able to Semiotics digest these huge numbers (in: Kirk (2012)). Perception Data Abstract visualization, in terms of its tone, is more about Communicate Process creating an aesthetic that portrays a general story or sense Aspects of Data Data Preparation of pattern Focus Representation You might not be able to pick out every data point or Presentation category, but there is enough visual information to give you References a feel for the physicality of the data This next image visualizes the global airline transportation network The project was designed to assess the threat of infectious diseases SoSe 2017 Jörg Cassens – Summary 66 / 180

Rehash Introduction History Semiotics Perception Intent: The Visualization’s Function Data Explain Communicate Process Explore Aspects of Data Exhibit Data Preparation Focus Intent: The Visualization’s Tone Representation Pragmatic and analytical Presentation Emotive and abstract References SoSe 2017 Jörg Cassens – Summary 67 / 180

Topic Introduction History Semiotics Perception Data Communicate Process Process Aspects of Data Data Preparation Focus Representation Presentation References SoSe 2017 Jörg Cassens – Summary 68 / 180

Process: Fry Acquire: Obtain the data, whether from a file on a disk or a Introduction source over a network. History Parse: Provide some structure for the data’s meaning, and Semiotics order it into categories. Perception Filter: Remove all but the data of interest. Data Communicate Mine: Apply methods from statistics or data mining as a Process Aspects of Data way to discern patterns or place the data in mathematical Data Preparation Focus context. Representation Represent: Choose a basic visual model, such as a bar Presentation graph, list, or tree. References Refine: Improve the basic representation to make it clearer and more visually engaging. Interact: Add methods for manipulating the data or controlling what features are visible. Source: Fry (2008) SoSe 2017 Jörg Cassens – Summary 69 / 180

Process: Yau Introduction What data do you have? History What do you want to know about your data? Semiotics What visualization methods should you use? Perception What do you see and does it makes sense? Data Communicate Process Aspects of Data Data Preparation Focus Representation Presentation References Source: Yau (2013) SoSe 2017 Jörg Cassens – Summary 70 / 180

Topic Introduction History Semiotics Perception Data Communicate Process Aspects of Data Aspects of Data Data Preparation Focus Representation Presentation References SoSe 2017 Jörg Cassens – Summary 71 / 180

Rooted in Data Introduction History Rooted in Data Semiotics “Visualization is ofen thought of as an exercise in graphic Perception design or a brute-force computer science problem, but the best Data Communicate work is always rooted in data. To visualize data, you must Process Aspects of Data understand what it is , what it represents in the real world , and in Data Preparation Focus what context you should interpret it in. Data comes in different Representation shapes and sizes, at various granularities, and with uncertainty Presentation attached, which means totals, averages, and medians are only a References small part of what a data point is about. It twists. It turns. It fluctuates. It can be personal, and even poetic. As a result, you can find visualization in many forms.” (Yau, 2013) SoSe 2017 Jörg Cassens – Summary 72 / 180

Time-Mapping Introduction A look at crashes over time shifs History focus to the events themselves Semiotics This Figure shows the number of Perception accidents per year, which tells a Data different story than the total Communicate Process seen before Aspects of Data Data Preparation Focus Accidents still occurred in the Representation tens of thousands annually, but Presentation there was a significant decline References from 2006 through 2010, and fatalities per 100 million vehicle miles traveled (not shown) also decreased Source: Yau (2013) SoSe 2017 Jörg Cassens – Summary 73 / 180

Granularity Seasonal cycles become obvious at month-by-month Introduction granularity, as shown in the next figure History Incidents peak during the summer months when people go Semiotics on vacation and spend more time outside, whereas Perception during the winter, fewer people drive, so there are fewer Data Communicate crashes Process Aspects of Data This happens every year Data Preparation Focus At the same time, you can still see the annual decline Representation overall between 2006 and 2010. Presentation References Source: Yau (2013) SoSe 2017 Jörg Cassens – Summary 74 / 180

Zooming In You can increase granularity to crashes by the hour Introduction The next figure breaks it down History Each row represents a year, so each cell in the grid shows Semiotics an hourly time series for the corresponding month Perception With the exception of a new year’s spike during the Data midnight hour, it’s hard to make out patterns at this level Communicate Process because of the variability Aspects of Data Data Preparation Actually, the monthly chart is hard to interpret, too, if you Focus don’t know what you’re looking for Representation Presentation References Source: Yau (2013) SoSe 2017 Jörg Cassens – Summary 75 / 180

Aggregation There are clear patterns, though, if you aggregate, as Introduction shown in the next figures History Instead of showing values at every hour, day, or month, you Semiotics can aggregate on specific time segments to explore the Perception distributions Data Communicate What was hard to discern, or looked like noise before, is Process Aspects of Data easy to see here Data Preparation Focus Representation Presentation References Source: Yau (2013) SoSe 2017 Jörg Cassens – Summary 76 / 180

Uncertainty There are a lot of examples for data with uncertainty Introduction Weather reports History Time to complete a file transfer Semiotics Remaining battery time Perception Data When you have data that is a series of means and medians Communicate or a collection of estimates based on a sample population, Process Aspects of Data you should always wonder about the uncertainty Data Preparation Focus Representation Example Presentation References The United States Census Bureau releases data about the country on topics such as migration, poverty, and housing, which are estimates based on samples from the population. A margin of error is provided with each estimate, which means that the actual count or percentage is likely within a given range SoSe 2017 Jörg Cassens – Summary 77 / 180

Sample Size Introduction History Semiotics Perception Data Communicate Process Aspects of Data Data Preparation Focus Representation Presentation Source: Yau (2013) References Uncertainty in statistical data can be reduced by using an appropriate sample size The needed sample size for a target uncertainty can be computed SoSe 2017 Jörg Cassens – Summary 78 / 180

Context makes Data Useful Introduction History Without context, data is useless, and any visualization you Semiotics Perception create with it will also be useless Data Using data without knowing anything about it, other than Communicate the values themselves, is like hearing an abridged quote Process Aspects of Data secondhand and then citing it as a main discussion point in Data Preparation Focus an essay Representation It might be okay, but you risk finding out later that the Presentation speaker meant the opposite of what you thought References You have to know the metadata, or the data about the data, before you can know what the numbers are actually about SoSe 2017 Jörg Cassens – Summary 79 / 180

Questions Introduction History Semiotics Who Perception collects the data Data is the data about Communicate Process How was it collected Aspects of Data Data Preparation What was collected Focus Representation When was it collected Presentation Where was it collected References Why was the data collected SoSe 2017 Jörg Cassens – Summary 80 / 180

Ethical Questions Is it always OK to make a visualization? Introduction Consider the following case History In 2010, Gawker Media, which runs large blogs like Semiotics Lifehacker and Gizmodo, was cracked, and 1.3 million Perception usernames and passwords were leaked Data They were downloadable via BitTorrent Communicate Process The passwords were encrypted, but the attackers cracked Aspects of Data Data Preparation about 188,000 of them, which exposed more than 91,000 Focus unique passwords Representation What would you do with that kind of data? Presentation References SoSe 2017 Jörg Cassens – Summary 81 / 180

What before How Introduction Next question: what is it we are trying to say with the History visualization we are developing? Semiotics We first need to determine what are the specific messages Perception we are looking to communicate to our audience – the what Data Communicate The how this is said will be covered in the design stage Process Aspects of Data This is roughly equivalent to a user-centred design process: Data Preparation before we look at how the application looks like, we first Focus Representation need to understand what the application should offer to Presentation the user References Editorial focus : An editorial approach to visualization design requires us to take responsibility to filter out the noise from the signals and to identify the most valuable, most striking, or most relevant dimensions of the subject matter SoSe 2017 Jörg Cassens – Summary 82 / 180

Topic Introduction History Semiotics Perception Data Communicate Process Data Preparation Aspects of Data Data Preparation Focus Representation Presentation References SoSe 2017 Jörg Cassens – Summary 83 / 180

Steps Introduction History Semiotics Acquisition Perception Examination Data Completeness Communicate Process Quality Aspects of Data Data Preparation Data Types Focus Representation Transformation Presentation For Quality References For Analysis Consolidation SoSe 2017 Jörg Cassens – Summary 84 / 180

Acquisition Introduction First, you need to get hold of your data History As discussed, this might already be provided to you from Semiotics those commissioning the work Perception You might have independently formed a sense of the Data Communicate specific subject dimensions on which you require data Process Aspects of Data Alternatively, it may be that you have yet to focus beyond a Data Preparation Focus broad subject level Representation Obtained from a colleague, client, or other third-party Presentation entity References A download taken from an organizational system Manually gathered and recorded Extracted from a web-based API Scraped from a website Extracted from a Documents (such as PDF files) SoSe 2017 Jörg Cassens – Summary 85 / 180

Examination Introduction History Once we’ve got the data, a thorough examination will Semiotics determine your level of confidence in the suitability of what Perception you have acquired Data Communicate This involves assessing the completeness and fitness of the Process Aspects of Data data to potentially serve your needs Data Preparation Focus will enable you to quickly scan, filter, sort, and search Representation through your data set Presentation Potential issues: References Completeness Quality SoSe 2017 Jörg Cassens – Summary 86 / 180

Examination: Completeness Introduction History Semiotics Perception Is it all there or do you need more? Data Is the size and shape consistent with your expectations? Communicate Process Does it have all the categories you were expecting? Aspects of Data Data Preparation Focus Does it cover the time period you wanted? Representation Are all the fields or variables included? Presentation Does it contain the expected number of records? References SoSe 2017 Jörg Cassens – Summary 87 / 180

Examination: Quality Introduction History Semiotics Are there noticeable errors? Perception Are there any unexplained classifications or coding? Data Communicate Process Any formatting issues such as unusual dates, ASCII Aspects of Data characters? Data Preparation Focus Are there any incomplete or missing items? Representation Presentation Any duplicates? Does the accuracy of the data appear fine? References Are there any unusual values or obvious outliers? SoSe 2017 Jörg Cassens – Summary 88 / 180

Data types: Categories Introduction History Semiotics Perception Data Categorical nominal Countries, gender, text Communicate Categorical ordinal Olympic medals, “Likert” scale Process Aspects of Data Quantitative (interval-scale) Dates, temperature Data Preparation Focus Quantitative (ratio-scale) Prices, age, distance Representation Presentation References SoSe 2017 Jörg Cassens – Summary 89 / 180

Data types: Operations N – Nominal (labels) Fruits: Apples, oranges, ... Introduction Operations: = � = History Semiotics O – Ordered ECTS Grades A, B, C, ... Perception Operations: = � = <> ≤≥ Data Communicate Process Q – Interval Dates: 19. Jan 2017 Aspects of Data Data Preparation (location of 0 arbitrary) Loc.: (LAT 33.98, LON -118.45) Focus Operations: = � = <> ≤≥ − Representation Like a geometric point. Cannot compare directly. Presentation Only differences (i.e. intervals) may be compared. References Q – Ratio Measurements: Length, Temp, ... (location of 0 fixed) Counts and amounts Operations: = � = <> ≤≥ −÷ Like a geometric vector, origin is meaningful. SoSe 2017 Jörg Cassens – Summary 90 / 180

Transformation for Quality Introduction History Semiotics This task is about tidying and cleaning your data in Perception response to the examination stage above Data We are looking to resolve any of the errors we discovered in Communicate Process order to transform the condition of the data we’re going to Aspects of Data Data Preparation be working with for our design Focus Representation Plugging the gaps caused by missing data, removing Presentation duplicates, cleaning up erroneous values, and handling References uncommon characters are some of the treatments we may be required to apply SoSe 2017 Jörg Cassens – Summary 91 / 180

Transformation for Analysis Introduction Here, we focus on preparing and refining it in anticipation History of its intended use for analysis and presentation Semiotics Parsing (split up) any variables, such as extracting year from Perception a date value Data Merging variables to form new ones, such as creating a Communicate Process whole name out of title, forename, and surname Aspects of Data Data Preparation Converting qualitative data/free-text into coded values or Focus keywords Representation Deriving new values out of others, such as gender from title Presentation or a sentiment out of some qualitative data References Creating calculations for use in analysis, such as percentage proportions Removing redundant data for which you have no planned use (be careful) SoSe 2017 Jörg Cassens – Summary 92 / 180

Resolution: Choice Introduction History Full : Plotting all data available as individual data marks Semiotics Perception Filtered : Exclude records based on a certain criteria Data Aggregate : “Roll-up” the data by, for instance, month, year, Communicate Process or specific category Aspects of Data Data Preparation Sample : Apply (mathematical) selection rules to extract a Focus fraction of your potential data Representation Particularly useful during a design stage if you have very Presentation large amounts of data and want to quickly develop References mock-ups or test out ideas Headline : Just showing the overall statistical totals SoSe 2017 Jörg Cassens – Summary 93 / 180

Consolidation When you originally access your data, you will likely Introduction History believe, or hope that you have everything you need Semiotics However, it may be that afer the examination and Perception preparation work, you identify certain gaps in your subject Data matter Communicate Process Additional layers of data may be required to be combined Aspects of Data Data Preparation (“mashed-up”) with our existing dataset, applied to Focus Representation perform additional calculations, or just to sit alongside this Presentation initial resource to help contextualize and enhance the References scope of our communication Always spend a bit of time considering if there is anything else you anticipate needing to supplement your data to help frame the subject or tell the stories you want to communicate SoSe 2017 Jörg Cassens – Summary 94 / 180

Topic Introduction History Semiotics Perception Data Communicate Process Focus Aspects of Data Data Preparation Focus Representation Presentation References SoSe 2017 Jörg Cassens – Summary 95 / 180

Visual Design Options Introduction History The way that you choose to represent your data – your Semiotics selection of chart type – should be influenced by the Perception questions you are trying to answer Data Communicate If you are asking a chart to facilitate a comparison between Process Aspects of Data the values of different categories, you might deploy a bar Data Preparation Focus chart Representation You wouldn’t use a line chart, unless you wanted to show Presentation how a value or values change over time References A scatter plot can be the perfect method of comparing two quantitative values for different countries SoSe 2017 Jörg Cassens – Summary 96 / 180

Data Sketching Introduction Consider the potential of visualization for ourselves History Visually analyzing a data set, and employing both inductive Semiotics and deductive reasoning, enables us to learn more about Perception our subject by exploring a dataset from all directions Data Communicate Rather than just looking at data, we are using visualization Process Aspects of Data to actually see it, to find previously undiscoverable Data Preparation Focus properties of our raw material, to learn about its shape, Representation and the relationships that exists within Presentation data sketching or pre-production visualization References Using visualization techniques to become more intimate with our raw material and to start to form an understanding of what we might portray to others and how we might accomplish that SoSe 2017 Jörg Cassens – Summary 97 / 180

Exploration Dimensions Introduction History Semiotics Comparisons and proportions Perception E.g. using bar charts Data Trends and patterns Communicate Process E.g. using line charts Aspects of Data Data Preparation Relationships and connections Focus Representation E.g. using scatter plots Presentation References The chart types shown illustrative just a small section of the gallery of options we have to call upon SoSe 2017 Jörg Cassens – Summary 98 / 180

Deductive Approach Introduction History Semiotics Deductive reasoning involves confirming or finding Perception evidence to support specific ideas Data A deductive approach to defining your data questions will Communicate Process involve a certain predetermined sense of what stories Aspects of Data Data Preparation might be interesting, relevant, and potentially available Focus within your data Representation Presentation You are pursuing a curiosity by interrogating your data set References in order to substantiate your ideas of what may be the key story dimensions SoSe 2017 Jörg Cassens – Summary 99 / 180

Inductive Approach Introduction History Inductive reasoning is much more open-ended and Semiotics Perception exploratory Data We are not sure what the interesting stories might be Communicate Process We use analytical and visualization techniques to try and Aspects of Data Data Preparation unearth potentially interesting discoveries, forming Focus Representation different and evolving combinations of data questions Presentation We may end up with nothing, we may find plenty References Fundamentally, this is about using visual analysis to find stories SoSe 2017 Jörg Cassens – Summary 100 / 180