CS 585: Lecture 9, Data Visualization (Data Viz)

What and why of data visualization

Data visualization turns data into graphics so we can see patterns, communicate insights, and support decisions for any kind of dataset.

Single-variable (1-D) visualization

Simple charts like pies, histograms, bars, bubbles, and word clouds show distributions or relative importance of one variable, and good design dramatically improves their clarity.

Multivariate data

Great multivariate graphics (like Minard’s Napoleon chart) can encode many variables at once in a single coherent 2-D visualization.

Spatial data visualization

Mapping data (points, choropleths, heatmaps) onto geography reveals spatial patterns, supports planning, and is widely used for demographics, health, and retail layout.

Spatio-temporal data

Overlaying time on maps (e.g., hurricane tracks and forecast cones) shows evolving paths and uncertainty, often via animations or sequences of time-stamped maps.

Interactivity & animation

Interactive controls and animations (filter, drill-down, time sliders) let users explore data more deeply and see how patterns change over time.

Real-time visualization

Real-time dashboards (traffic, stocks, earthquakes, cyberattacks, etc.) show live data streams, adding immediacy and practical relevance.

Network visualization

Network diagrams use nodes and edges with visual encodings (size, color, thickness, direction) to reveal relationships and structure in connected systems.

Notable visualization examples

Famous examples (collaboration maps, language maps, shot charts, media bias plots, daily routine timelines) illustrate how diverse topics can be made intuitive through visual encoding.

How to generate data viz (tools)

A huge ecosystem of tools—from R/Python libraries and JavaScript (d3.js) to Tableau/Qlik and online chart builders—supports creating visualizations, and it’s valuable to learn several.

The “science” of data visualization

Data viz combines art and science, drawing on perception, color, design, semiotics, data types, graphic variables, and interaction patterns to match visualization forms (tables, charts, maps, networks) to analytic needs.

Good design & resources

Good visualizations focus attention on the data itself rather than the technology, and many courses and resources (Tufte, galleries, talks, newsletters) can help develop strong data-viz design skills.