Data Story Principles and Examples

Dr. Nathaniel Cline

Warm-Up Exercise

USAID agriculture infographic

1. What data sets are used?

2. How are the data presented?

3. What is the core story the infographic is trying to tell?

4. What works to tell this story and what doesn’t? Should anything be added or removed?

the USAID graphic was commissioned as part of a series for their 50th anniversary celebration, focused on highlighting success but also the challenges ahead. We’ve been told it was published online but also displayed in their entrance as a large banner. It was created for the USAID by an independent design firm named Lemonly.

1. what data? For the USAID graphic, these often include answers like “population data over time”, “crop yield predictions”, and “animal ownership in Ecuador”

2. how are data presented? For example, in the USAID graphic population size is represented by a globe increasing in size and percentages are represented as donut charts. Briefly share how this sets up a consistent set of charts and symbols and ways to use them. We call this the “visual vocabulary” of the story.

3. what story? In the USAID example most people think the story is something like “food needs are rising, and if we can help female farmers we’d be better prepared to produce enough food.”

4. effective? What elements best support the story? Are there extraneous pieces that don’t help the story along? Are there pieces that actively distract from the main point or contradict it? In the USAID example focus on the takeaway message - what is it? Is it emphasized enough? Does it solve the problem? Are there examples of data that could be removed from the USAID story to make it more effective?

Tell students that these critiques will become a routine part of the course, and we will get better at it as we go.

https://databasic.io/en/culture/deconstruct-a-dataviz

Agenda

1

The R Programming Language

2

Visualizing Data

3

Data, Pictures, and Stories

4

Review and to do

The R programming language

Tools available

There are a huge number of data visualization tools, why use R which has a relatively steep learning curve?

• Flexibility

• Industry standard

• Translation to other languages

Excel

This is (maybe?) familiar to you as a data format (rows and columns). Ask for a show of hands. This is called a tabular format.

Rstudio environment

We will be working in the programming language R

And using RStudio which layers on top of R

You will write code in the console (or actually in a markdown file but we will get to that).

You can look at your data in the viewer and it looks somewhat like Excel - but we will actually not be looking at our data in tabular form like this a lot. Instead we will be interacting with it in a very different way.

How many have seen R? How many have seen other programming languages?

Markdown

The environment we will work in will end up looking like this.

On the right side, things are a bit friendlier. There is nice formatting, the code is contained in little chunks surrounded by explanatory text.

On the left side there is no real formatting, it looks a bit less friendly, etc…

The left hand side is called a markdown document and it produces the document on the left hand side.

UNVotes Visualized

Our first exercise will use a data set on votes at the UN general assembly.

The particular votes are on:

• Arms control and disarmament
• Colonialism
• Economic development
• Human rights
• Nuclear weapons and nuclear material
• Palestinian conflict

Three countries: - Turkey - UK and NI - US

UNVotes Visualized

Y - axis is the percent of resolutions the country voted yes on

UNVotes Visualized

X - axis is the Year

UNVotes Visualized

Each square shows a different category of resolution.

UNVotes Visualized

We can see here an unsurprising result - which is generally good news as far as data goes!

UNVotes Visualized

How did we make this visualization?

The data came from three separate sources.

un_votes: how the countries voted un_roll_call: each voting session un_roll_call: linking roll call sessions to the particular issue

UNVotes Visualized

Then we wrote some code - no need to get wrapped up in this. But note that the three data sources appear earlier here in the code.

UNVotes Visualized

And we see the word join here, which indicates that these three data sets are being joined together

UNVotes Visualized

And then we filter for the particular countries we are interested in

UNVotes Visualized

ggplot is the command (and package) we are going to use to visualize data in R. You can see it has lots of options we will get to, but this command is telling R to make a graph.

UNVotes Visualized

And we can see the title and subtitle here

UNVotes Visualized

Here is the R-markdown environment that we will use and get familiar with. There are words (in blue) that then get rendered as text, and then code (in red) that gets run as code on the other side.

UNVotes Visualized

At the top here we can see the “meta-data” of the document that then gets rendered.

Your turn

You will be asked to work on this markdown document in posit.Cloud - a web based version of R-Studio.

Visit our RStudio site to start the project.

RStudio cloud will cost $5 a month to join.

First Assignment

Once you complete your registration for positCloud, you will be able to access the first assignment.

This will create a copy of the assignment for you and you can edit away.

Select all files, export them (name the file something convenient) and email it to me.

Visualizing Data

Cairo’s Principles

• Truthful
• Functional
• Beautiful
• Insightful
• Enlightening

Healey’s Principles

• Good aesthetics
• No substantive issues
• No perceptual issues
• Honesty and good judgement

Principles applied

• Let’s take a moment to put these principles into action

• We will return to our usual set of questions, but let’s first focus on apply Cairo and Healey to some visualizations

Class Critique

Class Critique

Class Critique

Pair and Share

Now its time to talk with eachother! Pair up with someone to discuss your visualization.

• introduce yourselves

• describe what you chose for your visualization (what is the data, form of presentation, and main story)

• focus on whether it is effective or not using Cairo and Healey

• After 5 min, switch

Pictures, and Narrative

Doesn’t the data just speak for itself?

• Why care about visualization, design, stories, or any of this?

• Isn’t this all too soft - shouldn’t we just rely on the hard facts?

Show me the data!

head(my_data, 10)

# A tibble: 10 × 2
       x     y
   <dbl> <dbl>
 1  55.4  97.2
 2  51.5  96.0
 3  46.2  94.5
 4  42.8  91.4
 5  40.8  88.3
 6  38.7  84.9
 7  35.6  79.9
 8  33.1  77.6
 9  29.0  74.5
10  26.2  71.4

mean(my_data$x)

[1] 54.26327

mean(my_data$y)

[1] 47.83225

cor(my_data$x, my_data$y)

[1] -0.06447185

oh no

Stories and facts

Facts

• scrutinize and fight facts we don’t like

• bend or break facts to support biases

• corrective facts can strengthen misinformation

• visualized facts are harder to reject

Stories

• engage more of the brain

• form a connection between storyteller and listener

• increase attention and empathy

• can reduce skepticism and increase openness

• enhance comprehension

Review and to do

Review

1

The R Programming Language

2

Visualizing Data

3

Data, Pictures, and Stories

To do

1

Read

Chapter 4 in Fundamentals of Data Visualization (Wilke 2018)

Summary of CRAP graphic design principles from Presentation Zen (Reynolds 2008)

Chapter 4 in Effective Data Storytelling (Dykes 2020)

2

Do

RStudio Cloud Primers

UNdata exercise