critique – Page 18 – Coffee Spoons

Southwest 1380

On Tuesday, Southwest Flight 1380 made an emergency landing here in Philadelphia after the Boeing 737-700’s port engine exploded. One passenger died, reportedly after being partially sucked out of the aircraft after the explosion broke a window. But the pilot managed to land the aircraft with only one engine and without any further deaths.

I wanted to take a look at some of the eventual graphics that would come out to visually explain the story. And as of Thursday, I have seen two: one from the Guardian and another from the New York Times.

The Guardian’s piece is the simpler of the two, but captures the key data. It locates the engine and the location of the window blown out by debris from the engine.

The Guardian's graphic — The Guardian’s graphic

The New York Times’ piece is a bit more complex (and accompanied elsewhere in the article by a route map). It shows the seat of the dead passenger and the approximate locations of other passengers who provided quotes detailing their experiences.

So the first thing that struck me was the complexity of the graphic. The Times opted for a three-dimension model whereas the Guardian went with a flat, two-dimensional schematic of the aircraft. Notice, though, that the seating layout is different.

Four rows ahead of the circled window location are two seats, likely an exit row, in the Guardian’s graphic where in the Times’ piece they have a full three-seat configuration. If you check seating charts—seatguru.com was the first site that came up in the Google for me—you can see that neither configuration actually matches what the seating chart says should be the layout for a 737-700. Instead it, the Guardian’s more closely resembles the 737-800 model.

Nerding out on aircraft, I know. But, it is an interesting example of looking at the details in the piece. The Guardian’s piece is far closer to the layout, as least as provided by SeatGuru, and the New York Times’ is more representative of a generic narrow-body aircraft.

Personally, I prefer the Guardian in this case because of its improved accuracy at that level of detail. Though, the New York Times does offer some nice context with the passenger quotes. Unfortunately, the three-dimensional model ultimately provides just a flavour of the story, compared to the drier, but more accurate, schematic depiction of the Guardian.

Credit for the Guardian piece goes to the Guardian’s graphics department.

Credit for the New York Times piece goes to Anjali Singhvi, Sahil Chinoy, and Yuliya Parshina-Kottas.

New York Is Still Beating London

So two weeks ago I posted about the graphics in a BBC article about how London has surpassed New York in terms of murders, due to a spate of stabbings in the British capital. Well, somehow I missed this: an article from the Economist that rebuts that point. And it does it brilliantly.

Lies, damned lies, and statistics.

I think everybody who works with data knows that adage. Now, I am not using it to say that the BBC—or the numerous other media outlets that ran the story—lied. Just that it is easy to change the story based on the data, how it is presented, or which subsets of the data are selected.

The Economist’s article points out that the surpassing of New York is a short term data point, a worrying short term trend, definitely, but they then look at the data. They select two timeframes and look at them side-by-side.

It's all about what data you show, choose to highlight, and then how you show it. — It’s all about what data you show, choose to highlight, and then how you show it.

And that is what I love about this piece. It shows the long-term context of New York having a far-higher medium-term history of murder (some 28 years of data is shown). When I was growing up in the 90s, murders in New York—and to be fair almost all large American cities—was just something that was a known fact. During that time, London hovered below 200 or so, compared to the 2000+ in early 90s New York.

But then they also show the short term, which does point to a steady rise in London murders. But, the data could also show a one-time dip in the murders in New York. But they also show that the total number of deaths is still higher in New York than London, despite the three months of data.

Murder is not good. But these graphics are a good example of how selecting different time series for the same data set, and then showing which parts of the data to show. The earlier BBC piece, and my revision of it, did not show the total deaths. Nor did either piece show the longer timeline of data.

Credit for the piece goes to the Economist graphics department.

News Deserts

Yesterday we looked at the shrinking Denver Post. Today we have a graphic from a related story via Politico. The article explores the idea that President Trump performs better in what the article terms “news deserts”, those counties with a very low level of newspaper circulation. (The article explains the methodology in detail.) This piece we are looking at here shows how those counties performed against the circulation rate and their 2016 presidential election result.

Overall, the work is solid. But I probably would have done a few things differently. First, the orange overlay falls in the middle of one column of dots. Do those dots then fall inside or outside the categorisation of news desert?

Secondly, the dots. If this were perhaps a scatter plot comparing the variables of circulation rates and, perhaps, election vote results as a percent, dots would be perfect. Here, however, they create this slightly distracting pattern in the the main area of counties. When the dots are stacked neatly and apart from other columns, as they are more often on the right, the dots are fine. But in the packed space on the left, not as much.

As I was reading through the article I had a couple of questions. For example, couldn’t the lack of newspapers be reflective of the urban–rural split or the education split, both of which can be seen in the same election results. Thankfully the article does spend time going through those points as well. It is a bit lengthy of a read—with a few other perfectly fine graphics—but well worth it.

Credit for the graphics goes to Jeremy C.F. Lin.

Deaths in America

Yesterday was murders in London and New York. Today, we have a nice article from FiveThirtyEight about deaths more broadly in America. If you recall, my point yesterday was that not all graphics need to be full column width. And this article takes that approach—some graphics are full width whereas others are not.

This screenshot shows a nice line chart that, while the graphic sits in the full column, the actual chart is only about half the width of the graphic. I think the only thing that does not sit well with me is the alignment of the chart below the header. I probably would align the two as it creates an odd spacing to the left of the chart. But I applaud the restraint from making the line charts full width, as it would mask the vertical change in the data set.

The screenshot is of the graphic's full width, note the lines only go a little over half the width. — The screenshot is of the graphic’s full width, note the lines only go a little over half the width.

Meanwhile, the article’s maps all sit in the full column. But my favourite graphic of the whole set sits at the very end of the piece. It examines respiratory deaths in a tabular format. But it makes a fantastic use of sparklines to show the trend leading towards the final number in the row.

Credit for the piece goes to Ella Koeze and Anna Maria Barry-Jester.

Visualisation Types

I laughed aloud when I read this. So I am just going to leave it here. Happy Friday, all.

Credit for the piece goes to Randall Munroe.

The Cost of the US Tax Cut

I know I’ve looked at the Times a few times this week, but before we get too far into the next week, I did want to show what they printed on Saturday.

It is not too often we get treated to data on the front page or even the section pages. But last Saturday we got just that in the Business Section. Two very large and prominent charts looked at federal government borrowing and the federal deficit. Both are set to grow in the future, largely due to the recently enacted tax cuts.

That's about half the page on those two charts. — That’s about half the page on those two charts.

The great thing about the graphic is just how in-the-face it puts the data. Do two charts with 14 data points (28 total) need to occupy half the page? No. But there is something about the brashness of the piece that I just love.

And then it continues and the rest of the article points, at more normal sizes, to treasury bill yields and car loan rates. The inside is what you would expect and does it well in single colour.

But I just loved that section page.

Credit for the piece goes to Matt Phillips.

Russia Tomorrow

In news that surprises absolutely nobody, Russia “re-elected” Vladimir Putin as president for another six-year term. The Economist recently looked at what they termed the Puteens, a generation of Russians born starting in 1999 who have no memory of a Russia pre-Vladimir Putin.

This piece features a set of interactive dot plots that capture survey results on a number of topics that are segmented by age. It attempts to capture the perspective of Puteens on a range of issues from their media diet to foreign policy outlook to civil rights.

The design is largely effective. The Puteen generation sticks out clearly as the bright red to the cool greys. And more importantly, when the dots would overlap they move vertically away from the line so users can clearly see all the dots. And on hover, all the dots of the same age cohort’s interest are highlighted. I think one area of improvement would have been to apply that same logic to the legend to allow the user to scroll through the whole dataset without always having to interact with the chart. But that is a minor bit on an otherwise really nice piece.

Credit for the piece goes to the Economist’s graphics department.

My Kingdom for a Needle

I am exhausted. I tried to stay up late enough to catch the absentee ballots from Washington County. Alas, I did not quite make it. (You better bet I will be drinking all the caffeine today.) But someone else did not quite make it through the night. Or rather, something. What was it? The New York Times election night needle.

To understand why the Times made the needle, read this really great explainer. The super short version: it tries to forecast the results of that particular election day, accounting for things like uncounted votes. On television, analysts and large interactive screens can show how, usually, urban districts are counted first then followed by slower-to-return rural areas. But for people following results solely online, those nuances might well be lost. Enter the needle.

Last night, like much of the Twitterverse I follow for politics, I had the needle open in one tab. But as the results began to come in, something odd was happening on the Times’ results page. The votes were being displayed in a precinct-by-precinct fashion in Allegheny, Washington, and Greene Counties. But Westmoreland was oddly grey. It turned out, the county elections board was not, I suppose, digitally publishing the precinct results, only county-wide.

Fun fact, the needle’s model is apparently built on precinct results. So how do you have a needle if something like 30% of the model’s required or expected data will not be available? The Times tweeted about it a few times, but ultimately pulled it down. Better to not have it and be right than have it wrong just to have it.

But that brings me to the second point about the needle. Well done to BuzzFeed’s Decision Desk HQ, who were presciently concerned about the ability of the county to get precinct level results up online. So they sent a reporter, as in a human being, to Westmoreland to get the analogue results and then upload them to BuzzFeed’s own results spreadsheet. (I never did find a BuzzFeed live results page.)

Who knows the budget difference between the New York Times’s graphics/politics desks and that of BuzzFeed’s, but the ability to put a single person in Westmoreland made the difference for BuzzFeed, whose coverage via Decision Desk HQ, made for a more compelling following because they were, old school like, reading out results as they came in via reporters. And because there were no exit polls in the election, we had to wait for all the votes. Strangely, it almost felt like watching a UK general election where you have to wait hours for some constituencies to announce results. Though this election had a noticeable lack of Raving Monster Loony candidates.

I bring up the BuzzFeed contribution to the night because it does show how sometimes the sheer fact of placing a reporter on the ground can yield tremendous results. Come November, no, I don’t think any single media organisation can afford to put a reporter in every single US county. But I would bet the Times will be working on how to better precinct proof their needle.

Credit for the piece goes to Nate Cohn, Josh Katz, Sarah Almukhtar, and Matthew Bloch.

Italian Election Results

Europe enjoyed some significant political news yesterday. First, Angela Merkel will serve a fourth term as chancellor as the SPD members voted to allow their party to enter into a grand coalition with Merkel’s CDU/CSU party.

But the more important story is that of the Italian elections, where the centre-left under Matteo Renzi was attempting a comeback against the populist parties the 5-Star Movement and the League, the latter an anti-immigrant party. Also in the mix was Silvio Berlusconi, whose Forza party won 14% of the vote and as a member of a right/centre-right bloc that won 37%.

So I chose to highlight the homepage of IL Sore 24 Ore, an Italian newspaper, that had the results displayed clearly.

Live results at the top of the home page

Of course the big problem is that I can neither read nor speak Italian. So figuring out just what every label is proved to be a bit tricky. But once you figure it out, it is quite clear. The nice blue banner for the real-time results (again with the assuming of translation) does a nice job of clearly separating itself from the rest of the page, but the tables inside are quiet and not screaming for attention. Instead the user is allowed to find his or her party of bloc of interest and then scan to the right for the bold number of seats in the respective chamber.

The results page is similarly nice, using clean and simple tables to organise the information. Using the Chamber of Deputies page as an example, the overall results appear on the left while important context via maps and specific regions appear to the right. All the while the use of simple typography and whitespace guide the user to the appropriate data set.

These are the results for the Chamber of Deputies, the equivalent to the US House or House of Commons

And lastly a screenshot of an article about the election results, none of which I can read. Here, instead of an interactive table or graphic, we have a static graphic showing the results. It certainly captures the results in this particular moment—exact seat numbers have not yet been released—but could grow stale as the day goes on. Although there very well could be a page with interactive results like this, but that I cannot find because, again, I cannot read Italian.

The centre-right bloc did well, as did the 5-Star Movement (M5S)

The design of the graphic is nice. It uses the popular half-circle arc to show who “crosses the finish line” in terms of blocs seating more than 50% of the chamber. But once again, I am most impressed by the clarity of the table and information displays through white space and typography. (Though I feel in this case white space should be more like light salmon-coloured space.)

Overall, the designers did a fantastic job of presenting the data and information, so well that a non-Italian could even figure it out.

Credit for the piece goes to the Il Sore 24 Ore graphics department.

Post-Brexit Trading

Off of yesterday’s piece looking at the potential slowdown in British economic growth post-Brexit, I wanted to look at a piece from the Economist exploring the state of the UK’s current trade deals.

I understand what is going on, with the size of the bubbles relating to British exports and the colour to the depth of the free trade deal, i.e. how complex, thorough, and wide-ranging. But the grouping by quadrant?

With trade, geographical proximity is a factor. Things that come from farther cost more because fuel, labour time, &c. One of the advantages the UK currently has is the presence of a massive market on its doorstep with which it already has tariff- and customs-less trade—the European Union.

Consequently, could the graphic somehow incorporate the element of distance? The problem would be how to account for routes, modes of transport, time—how long does a lorry have to queue at the border, for example. Alas, I do not have a great answer.

Regardless of my concepts, this piece does show how the most valuable trade partners already enjoy the deepest and largest trade deals, all through the European Union. And so the UK will need to work to replicate those deals with all of these various countries.

Credit for the piece goes the Economist Data Team.