Agenda

whoami

A few class policies

• Be kind

• Be understanding and have patience, with others and yourself

• Help others whenever possible

Truly the most important part of this class. Important not just in terms of decency, but also in your learning, and most importantly, for equity.

A bit more specific

Normally I would have information here about welcoming kids into class.

Because we're virtual, that part is both easier and harder.

If you need to not attend class, or a portion of class, for any reason, that is fine.

Ideally you would let me know ahead of time. But we're in the middle of a pandemic and life is cray. Please try to contact me beforehand. If this isn't possible, please check in with me after.

Last intro thing

• I'm here for you

• We won't have specific office hours, but know I'm always willing to meet

• This course, like all in the sequence, can be difficult. Don't suffer in silence. Don't do this alone.

Course Website(s)

Materials

• Nearly everything will be distributed through the repo and through the website.

• Please clone the repo now, if you haven't already. Pull each week for the most recent changes.

• We'll use Canvas for grading, and that is essentially it.

R Markdown notes

• These slides were produced with {xaringan}, an R Markdown variant. I encourage you to try it out and use it for your final project presentation.

• The website was also produced with R Markdown (sort of)

  • It's a {blogdown} website with some custom CSS and Hugo shortcodes

• This course is not just about data viz, but also mediums for communication. This includes websites and data dashboards among other possibilities.

I assume you

• Understand the R package ecosystem (how to find, install, load, and learn about them)

• Can read "flat" (i.e., rectangular) datasets into R

  • I don't care what you use, but you should be using RStudio Projects & the {here} package
  • See Jenny Bryan's blog post for why.

• Can perform basic data wrangling and transformations in R, using the tidyverse

  • Leverage appropriate functions for introductory data science tasks (pipeline)
  • "clean up" the dataset using scripts and reproducible workflows

• Use version control with R via git and GitHub

• Use R Markdown to create reproducible dynamic reports

Learning objectives

• Transform data in a variety of ways to create effective data visualizations

• Understand and fluently apply different types of data joins

• Understand best practices in data visualization

• Customize ggplot2 graphics by reordering factors, creating themes, and applying ggthemes

• Create an online data visualization portfolio using distill and/or flexdashboards to demonstrate key learning

Examples

Below are some links to final projects from students who took this class last year.

Weekly learning objectives

Provide you a frame for what you should be working to learn for that specific week.

This week's objectives

• Understand the requirements of the course
• Understand the requirements of the final project
• Be ready to go with git and GitHub

Required Textbooks (free)

Healy

Wilke

Other books (also free)

Some options

• Geographic data (we'll have an intro, but there's a ton here and we won't really do it justice)
• Network data
• Text data
• DAGs
• Flow data (e.g., alluvial diagrams)
• Relational data (SQL & friends)
• Interactive plots
• Animated plots

Timo Grossenbacher

Labs

See the assignments page of the website.

10 points each (30 points total; 15%)

1. Distributions & GitHub collabo

2. Visual perception & plot reproducing

3. Color

Homework

20 points each (40 points; 20%)

• Basically the same as the labs, but scored correct/incorrect, and no in-class time devoted to them.

• Okay to work on collaboratively - I actively encourage you to do so as long as you're using a shared repo

• Homework 1: Creating new visuals while utilizing different types of joins

• Homework 2: Visualizing uncertainty, tables, and plot refinement

Quick note on reproducibility

A great blog post by Rafael Irizarry shows how almost any plot you see in popular media can be reproduced in R with ggplot.

For example

WSJ Version

ggplot reproduction

Data viz "in the wild" presentations

Everyone will present for ~5 minutes - order randomly assigned (coming up next)

• Find two data viz examples intended for two different audiences

• Discuss the following

  • What's trying to be communicated
  • How effective do you judge it? Why?
  • At least 1 area of strength
  • At least 1 area for (potential) improvement

Presentation order

I will email this out as well.

Six parts

• Proposal (10 points): Due 1/27/21

• Draft (15 points): Due 2/24/21

• Peer review (15 points): Assigned, 2/24/21; Due 3/3/21

• Presentation (20 points): 3/8/21 and 3/10/21 (Week 10)

• Product (60 points): Due 11:59:59 PM, 3/17/21

Product

Four components:

• A web-deployed portfolio showcasing your #dataviz skills.
  • distill (what I'll lecture on), R Markdown, or blogdown website
  • Technical document with pagedown or bookdown
  • Scientific poster with pagedown
  • flexdashboard

• At least three finalized data displays, with each accompanied by a strong narrative/story, as well as the history of how the visualization changed over time.

• Housed on GitHub

  • Fully reproducible

• Deployed through GitHub pages (or netlify or similar)

Proposal

Four components:

• Description of the data source (must be publicly available)

• Preliminary ideas of different viz

• Identification of the intended audience for each viz

• The intended message to be communicated for each viz

Draft

• Expected to still be a work in progress

  • Data visualizations should be largely complete

• Deployment not expected

• Provided to your peers so they can learn from you as much as you can learn from their feedback

Peer Review

• We are all professionals here. It is imperative we act like it.

• Understand the purpose of the exercise.

• Zero tolerance policy for inappropriate comments

• Should be vigorously encouraging

Utilizing GitHub

You'll be assigned three proposals to review (5 points each)

• Fork their repo, embed comments & suggest changes to their code, submit a PR

Presentation

Order randomly assigned. Basically a chance to share what you created!

• Discuss what is trying to be communicated

• Share the final products

• Discuss the progression along the way and why specific changes were made

Points

200 points total

• 3 labs at 10 points each (30 points; 15%)
• 2 homework assignments at 20 points each (40 points; 20%)
• five-minute data visualization "in the wild" presentation (10 points; 5%)
• Final Project (120 points; 60%)
  • Proposal (10 points; 5%)
  • Draft (15 points; 7.5%)
  • Peer review (25 points; 12.5%)
  • Presentation (10 points; 10%)
  • Product (60 points; 30%)

Grading

Demo

• The gitkraken GUI
• Creating a GitHub repo
• Sharing access (or creating an organization)
• Cloning the repo
• stage, commit, push
• pull
• branching
• forking and issues