e
A Great Example of Better Data Visualization: This Voting Map GIF
"Land doesn't vote. People do"
Done properly, the art of data visualization can be an incredibly powerful tool for educating people. It allows us to understand things that would otherwise be ungraspable due to their sheer complexity.
Done improperly, data visualization can be incredibly misleading. It's important that we have talented and hopefully unbiased (I know, what are the odds) designers presenting the information.
As one example of how bad data visualization can mislead, take a look at this map below. What you're seeing is a map of how each county in the United States voted in the 2016 Presidential election (Red = Republican, Blue = Democrat).
It looks like a landslide--because visually, it is. However, this is a wildly inaccurate representation of proportionality vis-à-vis the population, because all of those little shapes representing counties have vastly different amounts of people living within them. As some might put it, "Land doesn't vote. People do."
Data scientist Karim Douïeb figured that a more accurate way to represent how people voted is to use colored dots, varied in size proportionally to the population of each county. He turned the results into this GIF, which provides a clearer picture:
Pretty eye-opening, no? And yet, while this is clearly an improvement over the ham-fisted method of the first map in this entry, even this is not quite accurate. Within each of those large blue dots, you still have plenty of people who voted red, and vice versa. These results only show you which party won the vote in each region.
What do you think we'd see, if this data represented actual individual votes and we could zoom in on each one? The country is now more divided than ever, and just about evenly split. So all I'm certain of is that zooming out, we'd see a perfect shade of purple.
(Note: For those of you who'd like to further investigate or verify, Douïeb cites this "How well does population density predict U.S. voting outcomes?" article and analysis as the source of his data.)
-
o9Favorite This
-
Q21Comment
Directory Company Profiles
Core77
Industrial Design
Resources
Industrial Design
Resources
-
Core77 Directory
Find the Best Industrial Design CompaniesGo -
Goddard, Inc.
Goddard Inc. is a full-service engineering and industrial design firm specializing in the developmen...
-
Astro StudiosAstro Studios of California launched in 1994, and since then our multidisciplinary teams have design...
-
andesignlab
ANDESIGN is an industrial design consultancy based in Orange County California and Austin Texas. AN...
-
Open Design
We partner with teams to deliver fresh thinking and award winning products for passionate audiences....
-
Pushstart Creative
What is your future in a connected world? We create impactful strategies and build beautiful connec...
-
Anvil Studios, Inc
Anvil Studios is a product design firm for the real world. Anvil is Treasure Hinds and Greg Janky. ...
-
Endeavor Design
We help transform ideas into compelling products that grow your business.
K
{Welcome
Create a Core77 Account
Already have an account? Sign In
By creating a Core77 account you confirm that you accept the Terms of Use
K
Reset Password
Please enter your email and we will send an email to reset your password.
Comments
I have been a Core77 contributor and reader from the beginning. One thing that has always been off limits has been politics. Rain I have to say, this is a bit disappointing. I know the point of this post was data visualization, but it should be known that inadvertently would come out as a political conversation.
A Better Example of More Accurate Data Visualization: These Voting Cartograms from 2012 and 2016.
http://www-personal.umich.edu/~mejn/election/2012/
http://www-personal.umich.edu/~mejn/election/2016/
Luckily the author used the exact same page layout and images for both pages -- so if you want you can open the links in separate tabs and flip back and forth to see the difference between 12 and 16.
I personally would love to see the cartograms recreated by Karim Douïeb
That's great, except your whole premise "Land doesn't vote. People do." is flawed. One of the original reasons for having a bicameral house is that people vote through the House of Representatives and states or other smaller geographical based communities (analogous to land) votes through the Senate. And although the 17th Amendment superseded Article 1 Sec. 3, by replacing state appointment of senators with the popular vote (which by the way has had numerous negative unintended consequences, not the least of which is our national debt) the Electoral College and the duties of the Senate maintain the original governmental structure of our nation as a representative republic, not a democracy. This is important because it recognizes that not everything should be decided by a simple majority of the popular vote. What if everyone decided that slavery should be legal again? Would that make it right? The very word democracy tells us that with its root from the Greek demos, which contrary to popular belief doesn't mean "the people" so much as it means "idiot." Which is to say that sometimes, the majority is wrong. This is commonly referred to as Tyranny of the Masses. Now, make your jokes that the red regions are filled with idiots, but it was the blue regions that our system was designed to protect us from. Other, than that, cool graphic.
Gerald - I think you're missing the point. As a data visualization, the map doesn't accurately depict the data (i.e. the number of voters and how they're voting). The map represents land, so using it to depict election results is flawed. You entirely missed the point of the post.
Because Gerald, as the title of the thread says, people vote, not land and it is about the presidential election. The representative republic would survive if that election were a popular vote. Your entire post is pointless.
David, I am interested in discussing but not debating. I prefer civility over sarcasm. But thanks again for the response.
Except Gerald, the Senate serves as the representative republic. Presidential voting using the electoral system is not a prerequisite of a representative republic.
David, thanks for responding. I am not sure how either of your points refute mine, however, the fact that you're in Chicago and I'm in Southern Illinois seems to reflect our respective positions well.
there is a superfluous space
I like this very much. May I suggest for each circle; instead of winner takes all; replace them with a pie chart.
out of curiosity, can we get a visual of all those tiny spread out dots accumulated together? I mean, just bc there's land between them- a person's a person no matter how small ;)
Gosh, the snowflakes are falling.
Crazy, almost as if they’ve insulated themselves from other people’s ideas by creating echo chambers. Like if they were to say something and keep repeating it whether true or not it doesn’t matter.
Can you show me the counties that were less than 10% margin in purple.
Not many users are knowledgeable about the coding language or the words hence it is important that the data scientists know exactly how to create a visual form of the data curated so that when any user reads it they can understand what is being conveyed to them.
Here is a good article on this topic, along with a variety of examples:
And then there's XKCD's approach, showing actual stick figures for each 250K voters: https://xkcd.com/1939/
I have areas of question.
1. How many different red and blue blends (i.e., purples) can the typical human eye distinguish, especially when view a likely-compressed image on a small screen? Are these the best two colors to use?
2. At what level of zoom should there be a single blend that that represents multiple precincts? Should the viewer be able to control this, i.e., to toggle from viewing a dot for each precinct vs for each city, county, state?