critique – Page 29 – Coffee Spoons

England and Wales Census Results

Earlier this week, the Office of National Statistics in the United Kingdom released census results for England and Wales. (Northern Ireland and Scotland are reported separately.) England has more people than expected, most likely because of undercounting of immigrants, and Wales is now some three million and counting. There are fewer Christians than expected—and fewer Jedi than I expected—as the ranks of the non-religious grow. But from of course all of this comes a bevy of visualisations. These are but a few, but if anybody finds others worth nothing, please feel free to send them my way.

Straight from the source is a set of interactive mapping applications from ONS that compare 2001 data to 2011 data. As best it can, census districts are compared on a one-to-one basis, but with boundary changes that isn’t always possible. Clicking on district provides one with details about the responses for that area.

Perhaps the one thing missing from these—and it may well owe to the aforementioned boundary changes—is a map of changes to see which areas have been most impacted. Or a map of the results compared to the average to see where the average can be found and where the positive and negative extremes can be found.

An infographic from the Guardian looks at the overall dataset with quite a few maps and then circle-y things. While the large map is the white population in 2011, the remaining maps are before and after comparisons. Again, an interesting look would have been perhaps deviations from the average or of the actual change per district.

I appreciate the impact of the main story, the increasing diversity of England and, to a lesser extent, Wales. London in particular is now minority white. However, I am less keen on the circle-y things and that data could probably have been presented in a clearer, more direct fashion. I am not a fan of red, yellow, and green traffic light colours, but I also recognise that the Guardian is working within their brand on this.

Unfortunately this interactive map of Northern Ireland’s national identity does not quite work for me. I appreciate the toggle between the different response options, however, I find the responses themselves hard to compare. The colours remain the same, but the scales for the results change. For those identifying as Northern Irish, the top value is clearly less than those identifying as either British or Irish. But I would have liked to have seen the scales for British and Irish to closer match. I also find the black background distracting and overwhelming the colours. I wonder how the result would have worked if treated with the above aesthetic.

The BBC took a stab too with a section devoted to the results. Unlike the ONS visualisation above, however, the side-by-side comparison is forced to be smaller with the included text. And when one zooms into a particular district, the map degrades into crude polygons—a particular pet peeve of mine—that would be unrecognisable to someone familiar with the intimate geographic details of their home region. (Yes, simple shapes make the files smaller for overview maps, but when seen up-close, they lose their value by making ugly maps.) Also, the colours and bins in this particular view are not as informative as in the view above.

The BBC, however, did create a small graphic for an article that showed population changes in the districts, alas the colours did not work as well as one would hope.

That’s a lot for people to digest, but, overall I think the clearest visualisations go to the ONS. They lack the commentary that can be brought by journalism organisations, e.g. the BBC, but one needs a clear and powerful visualisation before one can start writing an analysis.

Credit for the ONS results goes to the ONS Data Visualisation Centre, for the Guardian infographic credit goes to Paul Scruton and Mark McCormick, for the Northern Ireland piece credit goes to John Burn-Murdoch, and credit for the BBC goes to the BBC.

Taxes

Let’s face it, governments need money to function. If you want a large military, you have to fund it. If you want pension system, you have to fund it. If you want medical care for the old, the sick, and the poor, you have to fund it. If you want to give everyone unicorns made of rainbow beams, you have to fund it. And…well…nevermind.

The point is taxes. After an election that focused so heavily on them, we’re still debating them. But here are some facts about them from the New York Times. The designers, Mike Bostock, Matthew Ericson, and Robert Gebeloff used small multiples of line charts—and lots of them—to look at who pays taxes by income band and how they pay different types of taxes. I found particularly interesting the points made near the bottom of the piece about how the progressive tax system is increasingly less so.

How the American tax system is becoming less progressive

But how do these taxes compare to spending? In a separate graphic for the same article, a stacked bar chart compares revenue to expenditure. With the exception of the balanced budget during President Clinton’s administration, we have been outspending our revenue since 1980. While statements to the effect of the US national budget needs to be managed like a US household budget are both overly simplistic and naive, there is a truth in a long-term mismatch between revenue and expenditure might cause problems. That is why many see the deficit and our debt as a medium-term problem facing the United States.

Credit for the first piece goes to Mike Bostock, Matthew Ericson, and Robert Gebeloff.

Maps Are Not Silver Bullets

I make a lot of maps in my line of work. Often times, they are not particularly interesting. Mostly because they follow similar patterns to this. More stuff is bought and sold where there are more people. More stuff is bought and sold where more people have more money. Et cetera, et cetera.

Maps are sometimes very useful. But I have a saying when people ask for a map of some kind of data tied to geographies: Maps are not silver bullets. That is to say, just because you throw data about countries, states, or counties onto a map does not mean you are going to see anything worthwhile let alone new or unexpected.

Credit for the XKCD piece goes to Randall Munroe

The Red Sox Were Truly Awful in 2012

The Boston Red Sox hired John Farrell this weekend to be their manager just one season after hiring Bobby Valentine for the role. There is a lot to be said about just who is to blame about the Red Sox’ awful season. But it was pretty awful. How awful? The Boston Globe shows us in this interactive piece.

It’s a series of small multiples of line charts. However, one of the big problems with the infographic is that the labels are entirely absent. As best I can tell the line is the number of games over .500, i.e. an even split between wins and losses. But, it could be more clearly called out if not in the legend or on the axes than in the title.

But over all it does put this past season into a sober perspective.

Credit for the piece goes to Daigo Fujiwara.

The Drop in the Unemployment Rate and How to Spin It

On Friday we received the monthly jobs report. And the furore that arose with it. Principally the anger stemmed from right-leaning commentators who believed that the non-partisan Bureau of Labor Statistics, a government agency tasked with collecting data on employment among other metrics, “cooked the books”/ “massaged the figures”/ flat-out lied to show a significant drop in the unemployment rate that could not be attributed to people who had stopped looking for work—a cause of some earlier drops over the last few years. As someone who works with data originally collected from national statistics offices across the world on a daily basis, those claims touched a nerve. But I shall leave that rant for another time.

Instead let’s look at the New York Times piece that quickly followed on the outrage of fools. We can look at and analyse the data in different ways—the origin of the phrase lies, damned lies, and statistics—and surely the Republican and Democratic parties would do just that. They did. This New York Times piece shows how that can be—and was—done. It involves points of reference and context.

First the facts:

Then how the Democrats spin them:

Finally how the Republicans spin them:

But the facts themselves do not lie. 114,000 non-farm jobs were added to payrolls. The unemployment rate fell to 7.8%, the lowest rate since January 2009.

Credit for the piece goes to Mike Bostock, Shan Carter, Amanda Cox and Kevin Quealy.

Better, Safer Driving

Hannah Fairfield at the New York Times created a great infographic a few years ago that looked at the history of the price of gasoline and how many miles, on average, an American drove in a car per year. The piece told some rather interesting stories starting in the 1950s with the explosion of the suburb, interstate highways, and car ownership. The energy crises of the late 1970s and early 1980s provided a spike that eventually subsided for the 1990s and early 2000s when the United States was the dominant economic power and the only country that really consumed that much gasoline. (I remember those days well for that was when I first started filling my own car’s gas tank. How great $1.xx/gal gas was.)

Earlier this week she returned in a similar fashion to look at driver safety over time. The metrics were average annual miles driven and the number of auto fatalities per 100,000 people. Segments of time characterised by a common theme, story, or technology are highlighted and the annotated to explain the change from the previous time period. It’s a rich story that walks the reader through the history of the American auto experience since World War II.

Credit for the piece goes to Hannah Fairfield.

The 47%

This is the year of the percentages. From 99 to 47. Earlier this week, Mother Jones revealed via a secretly recorded dinner Mitt Romney as claiming that he doesn’t care about the 47% of people who do not pay income tax. He probably meant that he doesn’t care about getting their vote rather than caring for them as people, but regardless of his intention his statement did not sound good. This 47% are people who are dependent on government, or as Romney’s vice presidential candidate would say, they are “takers”. But is this true?

The great thing about news stories and infographics is that as time progresses, one has more opportunities to find data to back stories and arguments. The day of the breaking news, CNN published this simple pie donut chart. Crude, but given the breaking news it is effective.

It proves that Romney was correct, that 47% of people do not pay income taxes. But, it also proves that Romney was at best glossing over the details or at worst manipulating the people listening into thinking that 47% of people do not pay taxes. It may be only 0.9% of Americans who do not pay any taxes. Furthermore, retirees and young people often do not earn income with which to pay taxes. If retired senior citizens are “takers”, I suppose Romney does not want the elderly vote so often important to the Republican base.

But news stories evolve and more statistics become available. A few days later, the New York Times published a longer infographic piece, with below a cropping of the overall.

It expands upon the nature of the story and breaks down the actual tax burden of the American public. It is far more nuanced that Romney stated back in May (the original recording of the video). The poorest Americans do not pay federal income taxes, they have the Earned Income Tax Credit (EITC) that is designed to deal with poverty, especially amongst families with children. It is, very simply, what one would call a tax credit incentivising work. The poorest Americans, however, still must pay payroll taxes and then state and local taxes. In percentage terms, the poorest Americans pay more to their states and local governments than do the wealthiest Americans, who in turn have the greater burden at the federal level. It is worth noting that many government programmes, local schools for example, are often funded at the state and local level.

The problem with Mitt Romney’s argument on taxes is that he wants to cut taxes at the highest income brackets and cut the social safety net programmes for the lowest brackets. We already have evidence that such policies do not correlate with economic growth. They instead correlate to a worsening gap between the wealthy and the poor. Think the 1920s rather than the 1950s, 60s, and 70s.

Tax policy is worth discussing. It is worth debating. But we should do so on the facts. I cannot recall the politician who spoke these words, but as someone once said, “you are entitled to your own opinion. You are not entitled to your own facts.”

Credit for the pieces go to Susie Poppick (CNN) and the New York Times.

Canadian Tradin’

The Globe and Mail of Canada published an infographic that where I work would probably be called a datagraphic. It presents data in a graphic fashion without a lot of context or conclusions that turn data into information. The piece in question looks at Canada’s balance of trade, i.e. how much it imports from other countries vs how much it exports to other countries.

While I appreciate the goal of the overall piece and fully understand that it may have in fact first lived in the print edition, the version shown on their website feels too large for the few data points contained within the graphic. The bars on the right and beneath the timeline are far too wide. The sections could likely have been condensed into a smaller, more compact space that would have given more visual weight to the timeline that clearly tells the story of a more volatile trading period for Canada since the global recession of 2008.

I also would probably change the chart type or simply look at a different data set for the trade balance with principal partners because the data for Japan barely registers. And while the other data can be seen, the minor differences are difficult to read. I would probably shift the emphasis from the actual dollar value of exports and imports to the percentage growth (or decline) of each over the last year.

Ivory Trade

This graphic from the New York Times looks at the illegal ivory trade out of Africa and into, primarily, the markets of Asia. I think the map works fairly well in showing why certain countries are centres for the illicit industry. But the two donut charts integrated into the graphic as part of the Indian Ocean are a bit weaker.

My main problem is that the shares are a bit difficult to distinguish as arcs, especially when looking at the export countries. But the second chart with the import markets does work a little bit better. In this case there are really only three markets: China, Thailand, and Others. But the chart contains the ambiguous China or Thailand. So in theory, that demarcation could fall anywhere between China and Thailand—a point harder made if comparing simply by bars. This means that the chart really is looking at China vs Thailand that combine to 87% vs. Others. The trick is finding the break between China and Thailand. Is this chart perfect? No, but in this case I think it an acceptable use of the donut—though I likely would have treated it a little bit differently to emphasise that point.

Crossing the Detroit River (Panic in Detroit)

The National Post’s business section, branded separately as the Financial Post, posted a comment about a proposed bridge that would span the Detroit River and add a third major crossing to the Detroit–Windsor area. The comment used a graphic to explain one of the key points of the story, that early 21st century traffic projections haven proven to be very much incorrect. Unfortunately, it took me a little bit of time to realise that in the graphic.

So without access to the raw data provided by United Research Services I have made a quick attempt to improve the graphic within the confines of Coffee Spoons’ main column space, i.e. 600 pixels. The original locator map is quite useful and therefore not included in my effort.

Detroit River Crossing Traffic Revisited

My main issues with the charts are the separation of the estimates from the actuals and the spacing between the estimates. I would have preferred to have seen, as in my example, how the actuals for 2010 fell far short of the 2004 projections. Ideally, I would have liked to have seen the original estimates for the intervening years between 2010, ’20, and ’30, however that data was not provided in the comment if it is even available from the original source. Consequently, unlike the original, I have kept the spacing of the actual data in the estimates with the intervening gaps.

The subtle effect of this increased spacing is to reduce the visual speed, if one will, of the projected growth. Over the original and narrower space the rate of increase appears fairly dramatic. However when given the correct spacing the ‘time’ to reach the projections lengthens and thus the rate ‘slows down’.

Credit for the original piece goes to Richard Johnson. The reinterpretation and any errors therein are entirely my own.