9 Data Visualizations Our Viz Team Loves

Published October 18, 2021

9 Data Visualizations our viz team loves

by IHME's Visualization Team

Every piece of data tells its own unique story. Data visualization requires developers to find a narrative in a set of data and translate those data insights into a clear, compelling narrative. 

The Institute for Health Metrics and Evaluation’s (IHME) Data Visualization team wears many hats to develop the visuals necessary to demonstrate population health data, thinking like data scientists, storytellers, and designers at the same time. The IHME Data Visualization team is composed of eight people who create and update about 15 visualization tools annually. These tools allow each user to utilize the interface for their own data visualization needs. For example, our GBD Compare tool alone can produce more than 3 billion estimates based on the user’s varied inputs. Our visualization tools allow policymakers, academics, and engaged members of the public to freely evaluate comparable measurements of the world’s most important health problems to fuel discussions and make informed decisions to best improve global health.

Recently, the IHME Data Visualization team has worked on Financing Global Health, Vaccine Hesitancy by Zip code, and COVID-19 visualization tools. The team works on producing visualizations on a variety of diverse topics and aspires to present them as creatively as possible. Where does the team’s inspiration for innovation come from? 

The following nine visualization tools are the team’s favorite visualizations that get their wheels turning andin their opinions—best express the stories of data. 

9. Karim Douïeb | United States 2020 Presidential Election Results

Image removed.

One of my favorite data visualizations from the past year is an animation created by Karim Douïeb. This visualization transitions between two views of the 2020 United States presidential election results: one looking at election results by landmass and the other by vote count. 

The initial view by landmass is overwhelmingly red, visually implying a Republican victory. But then the view changes, the geographical counties transform into circles, and the size of each circle begins to represent the total number of votes for a county. In this view you see the less populated areas recede and the more populated areas inflate. The outcome is a different view altogether, one that displays a truer representation of the results and reminds us that it isn't the landmass that is important, but the number of voters who inhabit those areas. 

This quick animation is a great reminder that in data visualization, how you choose to represent the data is sometimes more important than the data itself.

- Brian Dart

8. Bloomberg | measles Data vizImage removed.

One visualization that has inspired my work is the 2015 Bloomberg data story on measles outbreaks in the United States. What I love about this visualization is the way it leverages “object constancy” to guide the user through different dimensions of the data. In data visualization, “object constancy” refers to the practice of keeping data objects “constantly” in the view so a user can visually track the objects through animations. Object constancy decreases the cognitive burden associated with scanning labels or interpreting an entirely different chart, which makes a data story easier to follow. 

A brilliant example of this concept in the Bloomberg story is when the rectangles in a bar chart that represent measles cases per year break apart and recombine to form squares whose size represents measles cases by US state; we understand instantly that we’re transitioning from looking at measles cases by year from 2000 to 2015 to measles by US state over the same time period. Without that transition, we would need to read chart labels to understand what the next “view” represents, which breaks our train of thought as we’re trying to follow the larger story. We were so inspired by this that we implemented a very similar effect in the Child Mortality data story to transition from a map view to a simple dot plot of child mortality rates in Nigeria and Ethiopia.

- Ryan Shackleton 

7. Pitch interactive | Gun Violence

Image removed.

This data visualization by Pitch Interactive is a style of data visualization largely inspired by the critically acclaimed US Gun Killings data visualization (guns.periscopic.com). Mimicking the drone drop of explosives onto victims, this visualization does not shy away from communicating impact and emotion. It creates an animated story from the beginning of US drone strikes in 2004 and speeds up as it goes, illustrating the upsetting pattern and ubiquity of deaths caused by drone strikes.

The brilliance of the visualization technique is that it makes you feel sorrow at first for the individual and then horror at the collective death toll, and it does so without compromising its clarity or the detail of data that it displays. After the animation is complete, it allows the viewer to interact with the tool at will, exploring the specific instances as well as the data sources. It contextualizes the dataset and gives explanations in an understandable and accessible manner. 

- Kat Beame

6. Hanna Piotrowska | Analog Visualizations

Image removed.

I’ve selected a series of printed visualizations created by Hanna Piotrowska because they are beautiful and clearly communicate interesting data on a topic that we might not typically associate with a data visualization. She visualizes the content of the book If on a winter’s night a traveler, which has been described as a rumination on writing and reading, and as a challenging piece of conceptual art. We get to see Piotrowska’s chosen meta-analysis of the book’s content, and, in this way, I believe her work complements the unique intent of the novel itself. In particular, I love the sparseness of the visualizations. She leaves enough open space to allow your eye to focus on what’s important. And what’s important is made stark through the color of the data-carrying symbols. 

When a visualization is complex, Piotrowska includes a “How to read it?” foldout with a brief textual description of what is displayed and how it is displayed, as well as legends to describe each of the colors and symbols. And I think it’s very important to add this safety net, as it often makes the difference between a visualization being seen and understood or being seen and forgotten. It’s beautiful work that I hope to emulate—to at least some degree—in visualizations I work on in the future.

- Melissa Lafranchise

5. Our World In Data | Environmental impacts of food production

Image removed.

One recent data visualization I found compelling was Hannah Ritchie’s series on the Environmental Impacts of Food Production for Our World in Data. Through a mix of infographics and interactive charts accompanying her articles, she lays out the case that simple changes to our diets can have rather large impacts on global warming, land use, biodiversity, and more.

In one chart, she uses both color and iconography as preattentive attributes to anchor an association with the various stages of the food production supply chain. This is followed by a stacked bar chart summing up the total carbon emissions associated with many of the staple foods we eat, broken out by these supply chain stages. There is so much information embedded, but one key takeaway is that “eating local” has an almost negligible effect compared to overall food choice.

The charts and infographics each pack in a copious amount of data, yet are thoughtfully presented in a way that tells a story and lets the reader extract meaningful information. Rarely does a data visualization actually end up affecting my day-to-day decisions. This one did just that.

- Ben Hurst

4. Bloomberg2015 Was the Hottest Year On Record Image removed.

Originally published in early 2016 by Tom Randall and Blacki Migliozzi, this data story leaves you on the edge of your seat. Part of why this visualization has stuck with me over the years is that the animation walks you through a typical story arc. First, information is introduced, “here is what the average temperature used to be.” Next, the situation seems to improve when the average temperature decreases from 1890 to the 1930s. After that, the upward trend begins and you are on an elevator right to 2015, the hottest year on record up until that time.* I think the effective thing about this story arc is that there is no resolution: instead there is a call to action. 

Looking back at the version published in 2016, it is hard to miss the subtitle "Deny This," a reminder of how these types of visualizations are placed in the social context they are created in. 

*Editors’ note: The years 2016 through 2020 have all been hotter than 2014, with a general upward trend.

- Michael Fernandes

3. Charles Joseph Minard | Napoleon March Map

Image removed.

Visualization by: Charles Joseph Minard, Paris, November 20, 1869
Learn more: Wikipedia

The Napoleon March Map by Charles Joseph Minard is the number one visualization by Tableau, and I believe it is a very powerful one. In 1812, Napoleon marched to Moscow with plans of defeating the city. This map shows how disastrous it was. Starting with roughly 470,000 soldiers, Napoleon returned with just 10,000 remaining. This chart tells the story of that campaign and is one of the most famous visualizations of all time. 

The map details the out-and-back journey of Napoleon’s troops. The width of the line represents the total number of soldiers, and the color represents the direction (yellow for toward Moscow, black for the return trip). Below the central visualization is a simple temperature line graph illustrating the rapidly advancing winter cold. It is effective and well detailed, and shows different data points corresponding to the temperature, painting with just one simple color a staggering picture of the journey’s devastation.

- Maryam Zare

2.  THE NEW YORK TIMES | Who Gets to Breathe Clean Air in New Delhi? Image removed.

A recent visualization that really stood out to me was created by a collaboration between The New York Times and pollution researchers. The intentional choice to collaborate with in-country pollution researchers gives the authors a deeper understanding of the data, which translates into a compelling piece. “Who Gets to Breathe Clean Air in New Delhi?” is a data visualization accompanied by a story that pulls together minute-by-minute data and creates a narrative that guides readers through a day in the lives of two children.

My favorite part of this visualization is that it highlights the power of data storytellingalternating between powerful images and charts to form a cohesive narrative that is greater than the sum of its parts. The visualization evokes emotion throughout the piece by humanizing the data through imagery and storytelling, making the reader feel invested in the children’s lives.

Giving context to the data is crucial to tell a full story, especially when the visualization is focused on highlighting structural inequality. 

- Disha Patel 

1. THE NEW YORK TIMES | 512 Paths to the White House 

Image removed.

One of my favorite visualizations is the “512 Paths to the White House” created by Mike Bostock and Shan Carter for The New York Times. This visualization selects nine swing states from the 2012 election and illustrates the potential outcomes based on the winner of each state. Users can interact with the graph by selecting a winner for each state and seeing the potential remaining outcomes. This visualization does a good job of displaying a lot of information in a unique and intuitive way. I find interacting with the chart both fun and informative. It is a beautiful representation of a binary tree.

- Evan Laurie

Despite the fact that the data from these visuals shared are very different, they all accomplish the same goal of sharing a visually compelling story. Visuals make data accessible, digestible and can help remove the excess noise around data to highlight key points. Want to see some of the work that IHME’s team has developed? Check out the data visualizations here.