STAT 2430
Data Visualization

Course Mechanics

Andrew Irwin, a.irwin@dal.ca

Mathematics & Statistics,
Dalhousie University

2024-01-09

Syllabus

  • Goals

  • Resources

  • Evaluation

  • Weekly plan

Claus Wilke’s take on data visualization

Goals

  • Identify characteristics of a good data visualization.

  • Design, edit, and communicate with data visualizations.

  • Learn computer skills needed to create visualizations.

Learning outcomes

Visual impact and aesthetics of graphics

  • Connect data to perceptual features of your graphics.

  • Describe aesthetic features of good plots.

  • Use length, shape, size, colour, annotations, and other features to enable comparative visual interpretation.

Learning outcomes

Visualization as communication

  • Visualization as a tool for communication.
  • Iterative process similar to editing text.
  • Choose the right format for your audience and communications goals.
  • Evaluate effectiveness by peer-review and constructive feedback.

Learning outcomes

Computing skills

  • Create graphs with continuous and categorical variables with informative legends.
  • Add error bars, linear models, smooths, labels, and other annotations to a graph.
  • Use the principles of tidy data to help with transformation and analysis of data.
  • Summarize and transform data using dplyr.
  • Write reproducible reports to document your analysis process and present your results.
  • Distribute data, code, and results using git and github.
  • Develop skills for independently learning new data visualization methods and software.

Learning outcomes

Statistical models

  • Approximate data with LOESS, OLS, robust regression, polynomial regression.
  • Extract information from statistical models.
  • PCA to reduce the dimensionality of complex datasets.
  • MDS to visualize and compare similarities and dissimilarities between variables.
  • Divide observations into homogeneous and distinct groups using K-means.

Tools

  • R - software for statistics
  • Rstudio - a graphical interface to make using R easier
  • R markdown - write reports that combine computations, results, and natural language
  • ggplot - a tool for plotting data
  • git - a tool for managing your work and collaborating

Weekly Plan

  • Course notes online

  • In-class meetings

  • Evaluation

  • Office hours for individual discussion

Evaluation

  • Tasks (weekly)

  • Assignments (every two weeks)

  • Project (Proposal, Presentation, Report)

All work to be evaluated is described on a single course page.

Course Resources

  • Course web page

  • Course Notes

  • Textbooks

    • Healy. Data Visualization.
    • Wilke. Fundamentals of Data Visualization
    • Grolemund & Wickham. R for Data Science

  • Software: R, Rstudio, git

  • Brightspace