Election Day

The 2018 midterm elections are finally here. Thankfully for political nerds like myself, the New York Times homepage had a link to a guide of when what polls close (as early as 18.00 Eastern).

I'm not saying you can't keep voting. You just can't keep voting here.
I’m not saying you can’t keep voting. You just can’t keep voting here.

It makes use of small multiples to show when states close and then afterwards which states have closed and which remain open. It also features a really nice bar chart that looks at when we can expect results. Spoiler: it could very well be a late night.

But what I really wanted to look at was some of the modelling and forecasts. Let’s start with FiveThirtyEight, because back in 2016 they were one of the only outlets forecasting that Donald Trump had a shot—although they still forecast Hillary Clinton to win. They have a lot of tools to look at and for a number of different races: the Senate, the House, and state governorships. (To add further interest, each comes in three flavours: a lite model, the classic, and the deluxe. Super simply, it involves the number of variables and inputs going into the model.)

The Deluxe House model
The Deluxe House model

The above looks at the House race. The first thing I want to point out is the control on the left, outside the main content column. Here is where you can control which model you want to view. For the whimsical, it uses different burger illustrations. As a design decision, it’s an appropriate iconographic choice given the overall tone of the site. It is not something I would have been able to get away with in either place I have worked.

But the good stuff is to the right. The chart at the top shows the percentage of likelihood of a particular outcome. Because there are so many seats—435 are up for vote—every additional seat is between almost 0 and 3%. But taken in total, the 80% confidence band puts the likely Democratic vote tally at what those arrows at the bottom show. In this model that means picking up between 20 and 54 seats with a model median of 36. You will note that this 80% says 20 seats. The Democrats will need 23 to regain the majority. A working majority, however, will require quite a few more. This all goes to show just how hard it will be for the Democrats to gain a workable majority. (And I will spare you a review of the inherent difficulties faced by Democrats because of Republican gerrymandering after the 2010 election and census.) Keep in mind with FiveThirtyEight’s model that they had Trump with a 29% chance of victory on Election Day 2016. Probability and statistics say that just because something is unlikely, e.g. the Democrats gaining less than 20 seats (10% chance in this model), it does not mean it is impossible.

The cartogram below, however, is an interesting choice. Fundamentally I like it. As we established yesterday, geographically large rural districts dominate the traditional map. So here is a cartogram to make every district equal in size. This really lets us see all the urban and suburban districts. And, again, as we talked about yesterday, those suburban districts will be key to any hope of Democratic success. But with FiveThirtyEight’s design, compared to City Lab’s, I have one large quibble. Where are the states?

As a guy who loves geography, I can roughly place, for example, Kentucky. So once I do that I can find the Kentucky 6th, which will have a fascinating early closing race that could be a predictor of blue waviness. But where is Kentucky on the map? If you are not me, it might be difficult to tell. So compared to yesterday’s cartogram, the trade-off is that I can more easily see the data here, but in yesterday’s piece I could more readily find the district for which I wanted the data.

Over on the Senate side, where the Democrats face an even more uphill battle than in the House, the bar chart at the top is much clearer. You can see how each seat breakdown, because there are so fewer seats, has a higher percentage likelihood of success.

In the Senate, things don't look good for the Democrats
In the Senate, things don’t look good for the Democrats

The take away? Yeah, it looks like a bad night for the Democrats. The only question will be how bad does it go? A good night will basically be the vote split staying as it is today. A great night is that small chance—20%, again compared to Trump’s 29% in 2016—the Democrats narrowly flip the Senate.

Below the bar chart is a second graphic, a faux-cartogram with a hexagonal bar chart of sorts sitting above it. This shows the geographic distribution of the seats. And you can quickly understand why the Democrats will not do well. They are defending a lot more seats in competitive states than Republicans. And a lot of those seats are in states that Trump won decisively in 2016.

That's a lot of red states…
That’s a lot of red states…

I have some ideas about how this type of data could be displayed differently. But that will probably be a topic for another day. I do like, however, how those seats up for election are divided into their different categories.

Unfortunately my internet was down this morning and so I don’t have time to compare FiveThirtyEight to other sites. So let’s just wrap this up.

Overall, what this all means is that you need to go vote. Polls and modelling and guesswork is all for nought if nobody actually, you know, votes.

Credit for the poll closing time map goes to Astead W. Herndon and Jugal K. Patel.

Credit for the FiveThirtyEight goes to the FiveThirtyEight graphics department.

Congressional District Population Density

Tomorrow is Election Day here in the United States and this morning I wanted to look at a piece I’ve had in mind on doing from City Lab. I held off because it looks at the election and what better time to do it than right before the election.

Specifically, the article looks at the density of the different congressional districts across the United States. Whilst education level appears to be the most predictive attribute of today’s political climate—broadly speaking those with higher levels of formal education support the Democrats and those with lower or without tend to support President Trump—the growing urban–rural divide also works. But what about the in-between? The suburbs? The exurbs? And how do we then classify the congressional districts that include those lands.

For that purpose City Lab created its City Lab Congressional Density Index. Very simplistically it scores districts based on their mixture of low- to medium- to high-density neighbourhoods. But visually, which is where this blog is concerned, we get maps with six bins from pure urban to pure rural and all the mixtures in-between. This cartogram will show you.

All the urban and rural seats
All the urban and rural seats

Now, there are a couple of things I probably would have done differently in terms of the visualisation. But the more I look at this, one of those things would not be to design the hexagons to all fit together nicely. Instead, you get this giant gap right where the plains states begin west of the Mississippi River stretching through the Rockies over to California. If you think about it, however, that is a fairly accurate description of the population distribution of the United States. With a few exceptions, e.g. Denver, there are not many people living in that space. Four geographically enormous states—North Dakota, South Dakota, Montana, and Wyoming—have only one congressional district. Idaho has two. Nebraska three. And then Iowa and Kansas four. So why shouldn’t a map of the United States display the plains and Rocky Mountain interior as a giant people hole?

Like I said, initially I took umbrage at that design decision, but the more I thought about it, the more it made sense. But there are a few others with which I quibble. The labelling here is a big one. First, we have the district labels. They are small, because they have to be to fit within the five hexagons that define the districts’ shapes. But every label is black. Unfortunately, that makes it difficult to read the labels on the darker colours, most notably the dark purple. I probably would have switched out the black labels in those instances for white ones.

But then the state labels are white with black outlines, which makes it difficult to read on either dark or light backgrounds. The designer made the right decision in making the labels larger than the districts, but they need to be legible. For example, the labels of Alaska and Hawaii need not be white with black outlines. They could just be set in black type to be legible. Conversely, Florida’s, sitting atop darker purple districts, could be made white.

The piece makes use of more standard geographic map divided into congressional districts—the type you will see a lot tomorrow night. And it makes use of bar charts to describe the demographics of the various density types. I like the decision there to use a new colour to fill in the bars. They use a dark green because it can cut across each of the six types.

Credit for the piece goes to David H. Montgomery.

Which of These Countries Does Not Belong

For those of you reading from the States, I hope you all enjoyed your holiday. And for my UK readers, I hope you all enjoyed your summer bank holiday last weekend. So now to the good and uplifting kind of news.

Something is clearly not right here.
Something is clearly not right here.

Indeed, a chart about deaths from firearms from the Economist. From a graphical standpoint, we all know how much I loathe stacked bar charts and this shows why. It is difficult for the user to isolate and compare the profiles of certain types of firearm violence against each other. Clearly there are countries where suicide by gun is more prevalent than murder, but most on this list are more murder happy.

And then the line chart that is cleverly spaced within the overall graphic, well, it falls apart. There are too many lines highlighted. Instead, I would have separated these out into a separate chart, made larger, so that the reader can more easily discern which series belongs to which country. Or I would have gone with a set of small multiples isolating those nine countries.

I am also unclear on why certain countries were highlighted in the line chart. Did they all need to be highlighted? Why, for example, is Trinidad & Tobago. It is not mentioned in the article, nor is it in the stacked bar chart.

But the biggest problem I have is with the data itself. But, every one of the countries on that list is among the developing countries or the least developed countries. Except one. And that, of course, is the United States.

Credit for the piece goes to the Economist Data Team.

The Toll of the Trolls

This is an older piece that I’ve been thinking of posting. It comes from FiveThirtyEight and explores some of the data about Russian trolling in the lead up to, and shortly after, the US presidential election in 2016.

They're all just ugly trolls. Nobody loves them.
They’re all just ugly trolls. Nobody loves them.

The graphic makes a really nice use of small multiples. The screenshot above focuses on four types of trolling and fits that into the greyed out larger narrative of the overall timeline. You can see that graphic elsewhere in the article in its total glory.

From a design standpoint this is just one of those solid pieces that does things really well. I might have swapped the axes lines for a dotted pattern instead of the solid grey, though I know that seems to be FiveThirtyEight’s house style. Here it conflicts with the grey timeline. But that is far from a dealbreaker here.

Credit for the piece goes to Oliver Roeder.

The Freedom of the Press

By now you may have heard that this Thursday media outlets across the United, joined by some international outlets as well, have all published editorials about the importance of the freedom of the press and the dangers of the office of the President of the United States declaring unflattering but demonstrably true coverage “fake news”. And even more so, declaring journalists, especially those that are critical of the government, “enemies of the people”.

I have commented upon this in the past, so I will refrain from digressing too much, but the sort of open hostility towards objective reality from the president threatens the ability of a citizenry to engage in meaningful debates on public policy. Let us take the clearly controversial idea of gun control; it stirs passions on both sides of the debate. But, before we can have a debate on how much or how little to regulate guns we need to know the data on how many guns are out there, how many people own them, how many are used in crimes, in lethal crimes, are owned legally or illegally. That data, that verifiably true data exists. And it is upon those numbers we should be debating the best way to reduce the numbers of children massacred in American schools. But, this president and this administration, and certain elements of the citizenry refuse to acknowledge data and truth and instead invent their own. And in a world where 2+2=5, no longer 4, who is to say next that no, 2+2=6.

There are hundreds of editorials out there.

Read one from the Philadelphia Inquirer, the Chicago Tribune, the Guardian, and/or the New York Times.

But the one editorial board that started it is that of the Boston Globe. I was dreading how to tie this very important issue into my blog, which you all know tries to focus on data and design. As often as I stand upon my soap box, I try to keep this blog a little less soapy. Thankfully, the Globe incorporated data into their argument.

The end of their post concludes with a small interactive piece that presents survey data. It shows favourability and trustworthiness ratings for several media outlets broken out into their political leanings. The screenshot below is for the New York Times.

Clearly Republicans and Democrats view the Times differently
Clearly Republicans and Democrats view the Times differently

The design is simple and effective. The darker the red, the more people believe an outlet to be trustworthy and how favourably they view it.

But before wrapping up today’s post, I also want to share another bit from that same Boston Globe editorial. As some of you may know, George Orwell’s 1984 is one of my favourite books of all time. I watched part of a rambling speech by the president a few weeks ago and was struck at how similar his line was to a theme in that novel. I am glad the Globe caught it as well.

Credit for this piece goes to the Boston Globe design staff.

Radiohead in Philadelphia

A week and a half ago my favourite band, Radiohead, played two shows in Philadelphia to close out their 2018 North America tour. I got to see the final of the two shows. And I decided to make this little piece over the weekend. Because it was totally fantastic.

The data shows that the band played a good mix of songs from across their discography. Admittedly they played nothing from Pablo Honey, but with the exception of Creep and Anyone Can Play Guitar along with some of the era’s b-sides, I really do not listen to that album all that often. They also skipped over Amnesiac, but did play five songs from my favourite album, Kid A, so, yeah, again, totally fantastic. Especially those final three songs to close the main setlist. Just brilliant.

Two hours of amazing
Two hours of amazing

Credit for this work is mine.

Joblessness in the Developed World

  • We have been looking at tariffs a little bit this week, but unfortunately one of the side effects of tariffs is job losses. And of course when it comes to people losing jobs, not all countries in the  developed world handle them the same. Last month the Washington Post published an article examining how those countries compare in a number of related metrics such as unemployment compensation, notice for termination, and income inequality.
Not all countries give people the short stick.
Not all countries give people the short stick.

It uses a series of bar charts to show the dataset and reveal how the United States fares poorly compared to its peers. The chart above looks at the earning needed for termination from employment and the differences are stark. The outlined bar chart shows longer tenured employees and the full bars as coloured. Of course this makes it look like a stacked bar chart or filled bar chart. Instead I wonder if a dot plot would be clearer. It would eliminate the confusion in determining what if any share of the empty bar is held by the full bar.

The US offers shockingly little assistance to people
The US offers shockingly little assistance to people

The chart for unemployment insurance versus assistance is a bit better. Here the bar represents insurance and the lines assistance. I like how the lines continue off beyond the margins to indicate an unlimited timeframe for assistance. However, for those countries where assistance is short-lived, the bars versus lines again begin to look like an instance of a share of a total, which they are not.

Still a Loyalist

As most of you know, I am what would have been called a loyalist. That is, I disagree with the premise of the American Revolution. People often mistake that as saying I think Americans should be British. No, although I personally would not mind that. Instead, America would likely have been a lot more like Canada and it would have obtained its independence peacefully through an organic, evolutionary process leading to, likely, some kind of parliamentary democracy.

Every year, somebody digs up articles people have written about why the Revolution was a bad idea. I have seen a lot of them. But I had not seen this Washington Post article that looked at constitutional monarchies. It was published during the whole royal baby buzz back in 2013. It examines why constitutional monarchies are not so bad, and might even be better than presidential republics.

God save the Queen
God save the Queen

The above graphic is far from great. The same goes for the other graphic in the article. I probably would have added more emphasis on the constitutional monarchies as they get overwhelmed by the number of non-constitutional monarchies s in the scatter plot. That could be through a brighter blue or keeping the pink and setting the rest to a light grey. I perhaps would have added a trend line.

Credit for the piece goes to Dylan Matthews.

Going Over (But Actually Under)

Late last week I was explaining to someone in the pub why the World Cup matches are played beyond their 90 minute booking. For those among you that do not know, basically the referees add up all the stoppage time, i.e. when play stops for things like injuries or people dilly dallying, and then tack that on to the end of the match.

But it turns out that after I explained this, FiveThirtyEight published an article exploring just how accurate this stoppage time was compared to the amount of stopped time. Spoiler: not very.

In design terms, the big takeaway was the dataset of recorded minutes of actual play in all the matches theretofore. It captured everything but the activity totals where they broke down stoppage time into categories, e.g. injuries, video review, free kicks, &c. (How those broke out across an average game are a later graphic.)

Through 27 June
Through 27 June

The setup is straightforward: a table organises the data for every match. The little spark chart in the centre of the table is a nice touch that shows how much of the 90 minutes the ball was actually in play. The right side of the table might be a bit too crowded, and I probably would have given a bit more space particularly between the expected and actual stoppage times. On the whole, however, the table does its job in organising the data very well.

Now I just wonder how this would apply to a baseball or American football broadcast…

Credit for the piece goes to David Bunnell.

The World Cup Begins

If you live under a rock or in America, the World Cup starts today. (Go England.) So what else to have but a chart-driven piece from the BBC from last week about the World Cup. It features seven charts encapsulating the competition. But the one I want to focus on? It’s all about the host nations, in this case Russia.

To host, or not to host, that is the question of how much can you pay FIFA officials under the table…
To host, or not to host, that is the question of how much can you pay FIFA officials under the table…

On its design, I could go without the football icons to represent points on the dot plot, but I get it. (Though to be fair, they work well as icons depicting the particular World Cup event in another set of graphics elsewhere in the article.) In particular, I really like the decision to include the average difference between a host nation’s points in non-hosting matches vs. hosting matches.

It does look like the host nation scores more points per match than when they are not hosting. And that—shameless plug—reminds me of some work I did a few years back now looking at the Olympics and the host nation advantage in that global competition.

Credit for the piece goes to the BBC Data Team.