History of Data Viz

Communication with images. John W. Tukey. Its easier to understand concepts with images. Blueprints and Seismographs are data visualizations. Find patterns and develop hypothesis using the images. Communicate the results to others. Minard's Map of Napoleon's Diastrous RUssia 1812. Sankey diagram. John Snows 1854 Cholera Outbreak helped prove its water borne. Nightingale's Rose chart showed more soldiers died of disease.

Key Concepts

Is and Art and a Science. Data layer, mapping layer, graphic layer. Typical datasets have rows, and columns with variables. Nominal, ordinal, and quantitative variables.

Color

Hues, primary colors secondary colors. Saturation is purity of color.

Interaction Effects

Pointing with mouse. Touch. Region selection. Magic Lens. Zoom for better look. Filter to view subset. Extra detail on click.

Charts

Line chart changes over time. Bar chart compares categories. Radial: cool but misleading. Pareto descending bars with a line to show defects. Histogram, bins of data. Pie chart: parts of a whole. Sunburst chart shows a hierarchy. Charts can be intentionally misleading. Pie charts are the worst.

Geovisualization

Geospatial: data with geographic or location component. GeoJSON is the open standard. It specifies location, area, boundaries. GeoJson has Points(addresses,locations), lines (streets & boundaries), polygons(provinces & tracts). Longitude and Latitude. Origin is Gulf of Guinea. Map projections put the globe on screen but make it distorted. Mercator makes Europe look big. Tissot's indicatrix: grid of circles to show distortion. Gall Peters projection is more accurate in size but shapes are off. Winket-Tripel, Albers projections.

Topology: Relationshipsbetween points, lines, polygons that represent features or feature. Boundaries are shared arcs.

Type of maps: Choropleth: color each region. Bubble map: place bubbles with size based on value. Dot Map: put dots on the map to represent something. Heat Map: approximate locations to show something. Connection map: Arc between points. Hexbin Map: Represent locations as hexagons. Cartogram: distort map based on variable, artistic.

Big ol Chart List

Area chart: (line chart with fills underneath). Difference area chart. Stacked area chart. Stream graph. Scatter plot. Bubble chart. Radial Chart. Parallel coordinates. Waffle Chart. Word Cloud. Box and whisker chart. Violin plot. Candlestick chart. Gantt chart. Sankey diagram. Chord Diagram. Matrix diagram. Hierarchy chart. Line chart: changes over time. Bar chart: Compare categories. Pareto chart: Throw a line on there. Histogram: Plots distribution. Pie chart: Percentage of a whole, can be multilinear, sunburst for when levels mismatch. https://datavizcatalogue.com/

Score cards and dashboards

Color Theory

Can cause physical reactions. It is a visual responset to light reaching eye, an indirect physical propery. People have favorite colors. Warm & Cool colors make us feel different .Cool is relaxing, warm is stimulating. Red is like violence, passion, danager, importance. Raises blood pressure, in China it means prosperity. Orange is change and movement. Yellow is happiness and cowardice. Blue is calmness, responsibility, relaxing, and refreshing. Most people like blue; clean water and sky. Green is earthy and vibrant. Purple is royal. Women can see more colors.

Wave lengths determine color. White is a mixture of all colors. Black is body radiation? Human visual system uses Eyes, Rods+Cones. Color deficiency when cones are weak. There are different types of color blindness red-green, blue-yellow, monochromacy, ishiara color test.

Addative color mixing. Chromaticity Diagram. Subtractive color mixing. RGB Cube, color space in nonlinear. Color Wheel. Tint is color + White. Shade = Color + Black. Analogous colors, complementary colors.

Munsell color space. Munsells solid. Color brewer: Sequential darker hue for higher value. Divergent: Fuse in the middle with white.

Data Story Telling

Hans Rosling great presentation. Data, visuals, narrative. Explain whats happening, enlighten the audience, engage them.

Gestalt Laws

I Was late but from an hour or so in. Law if Proximity: things that are grouped. Law of simularity, beats proximity. Law of enclosure. Law of continuity. Law of connectivity: beats everything but enclosure. Metastability. Common fate: things that move together are related. Nearby shapes cause illusian and wreck our perception. Something slightly off draws our attention. Use this by greying things out. Figure/Ground: Help people differentiate between foreground and background.

Tufte's Principles

1. Graphical integrity. Tell the truth. Lie Factor = Graph Effect/Data effect. Use good labels, use correct number of dimensions. 2: Data-Ink ratio, use ink on actual data. 3: Chart junk, ditch it. 4: Data density, put a lot of info on the page. 5: Small multiples, series of related graphs.

spark lines: Small enough to be embedded in text.

Link's up!