A graphical compass

A Twitter user pointed me to this article from Washington Post, ruminating about the correlation between gas prices and measures of political sentiment (such as Biden's approval rating or right-track-wrong-track). As common in this genre, the analyst proclaims that he has found something "counter intuitive".

The declarative statement strikes me as odd. In the first two paragraphs, he said the data showed "as gas prices fell, American optimism rose. As prices rose, optimism fell... This seems counterintuitive."

I'm struggling to see what's counterintuitive. Aren't the data suggesting people like lower prices? Is that not what we think people like?

The centerpiece of the article concerns the correlation between metrics. "If two numbers move in concert, they can be depicted literally moving in concert. One goes up, the other moves either up or down consistently." That's a confused statement and he qualifies it by typing "That sort of thing."

He's reacting to the following scatter plot with lines. The Twitter user presumably found it hard to understand. Count me in.


Why is this chart difficult to grasp?

The biggest puzzle is: what differentiates those two lines? The red and the gray lines are not labelled. One would have to consult the article to learn that the gray line represents the "raw" data at weekly intervals. The red line is aggregated data at monthly intervals. In other words, each red dot is an average of 4 or 5 weekly data points. The red line is just a smoothed version of the gray line. Smoothed lines show the time trend better.

The next missing piece is the direction of time, which can only be inferred by reading the month labels on the red line. But the chart without the direction of time is like a map without a compass. Take this segment for example:


If time is running up to down, then approval ratings are increasing over time while gas prices are decreasing. If time is running down to up, then approval ratings are decreasing over time while gas prices are increasing. Exactly the opposite!

The labels on the red line are not sufficient. It's possible that time runs in the opposite direction on the gray line! We only exclude that possibility if we know that the red line is a smoothed version of the gray line.

This type of chart benefits from having a compass. Here's one:


It's useful for readers to know that the southeast direction is "good" (higher approval ratings, lower gas prices) while the northwest direction is "bad". Going back to the original chart, one can see that the metrics went in the "bad" direction at the start of the year and has reverted to a "good" direction since.


What does this chart really say? The author remarked that "correlation is not causation". "Just because Biden’s approval rose as prices dropped doesn’t mean prices caused the drop."

Here's an alternative: People have general sentiments. When they feel good, they respond more positively to polls, as in they rate everything more positively. The approval ratings are at least partially driven by this general sentiment. The same author apparently has another article saying that the right-track-wrong-track sentiment also moved in tandem with gas prices.

One issue with this type of scatter plot is that it always cues readers to make an incorrect assumption: that the outcome variables (approval rating) is solely - or predominantly - driven by the one factor being visualized (gas prices). This visual choice completely biases the reader's perception.

P.S. [11-11-22] The source of the submission was incorrectly attributed.

Painting the corner

Found an old one sitting in my folder. This came from the Wall Street Journal in 2018.

At first glance, the chart looks like a pretty decent effort.

The scatter plot shows Ebitda against market value, both measured in billions of dollars. The placement of the vertical axis title on the far side is a little unusual.

Ebitda is a measure of business profit (something for a different post on the sister blog: the "b" in Ebitda means "before", and allows management to paint a picture of profits without accounting for the entire cost of running the business). In the financial markets, the market value is claimed to represent a "fair" assessment of the value of the business. The ratio of the market value to Ebitda is known as the "Ebitda multiple", which describes the number of dollars the "market" places on each dollar of Ebitda profit earned by the company.

Almost all scatter plots suffer from xyopia: the chart form encourages readers to take an overly simplistic view in which the market cares about one and only one business metric (Ebitda). The reality is that the market value contains information about Ebitda plus lots of other factors, such as competitors, growth potential, etc.

Consider Alphabet vs AT&T. On this chart, both companies have about $50 billion in Ebitda profits. However, the market value of Alphabet (Google's mother company) is about four times higher than that of AT&T. This excess valuation has nothing to do with profitability but partly explained by the market's view that Google has greater growth potential.


Unusually, the desginer chose not to utilize the log scale. The right side of the following display is the same chart with a log horizontal axis.

The big market values are artificially pulled into the middle while the small values are plied apart. As one reads from left to right, the same amount of distance represents more and more dollars. While all data visualization books love log scales, I am not a big fan of it. That's because the human brain doesn't process spatial information this way. We don't tend to think in terms of continuously evolving scales. Thus, presenting the log view causes readers to underestimate large values and overestimate small differences.

Now let's get to the main interest of this chart. Notice the bar chart shown on the top right, which by itself is very strange. The colors of the bar chart is coordinated with those on the scatter plot, as the colors divide the companies into two groups; "media" companies (old, red), and tech companies (new, orange).

Scratch that. Netflix is found in the scatter plot but with a red color while AT&T and Verizon appear on the scatter plot as orange dots. So it appears that the colors mean different things on different plots. As far as I could tell, on the scatter plot, the orange dots are companies with over $30 billion in Ebitda profits.

At this point, you may have noticed the stray orange dot. Look carefully at the top right corner, above the bar chart, and you'll find the orange dot representing Apple. It is by far the most important datum, the company that has the greatest market value and the largest Ebitda.

I'm not sure burying Apple in the corner was a feature or a bug. It really makes little sense to insert the bar chart where it is, creating a gulf between Apple and the rest of the companies. This placement draws the most attention away from the datum that demands the most attention.




Finding the right context to interpret household energy data

Bloomberg_energybillBloomberg's recent article on surging UK household energy costs, projected over this winter, contains data about which I have long been intrigued: how much energy does different household items consume?

A twitter follower alerted me to this chart, and she found it informative.

If the goal is to pick out the appliances and estimate the cost of running them, the chart serves its purpose. Because the entire set of data is printed, a data table would have done equally well.

I learned that the mobile phone costs almost nothing to charge: 1 pence for six hours of charging, which is deemed a "single use" which seems double what a full charge requires. The games console costs 14 pence for a "single use" of two hours. That might be an underestimate of how much time gamers spend gaming each day.


Understanding the design of the chart needs a bit more effort. Each appliance is measured by two metrics: the number of hours considered to be "single use", and a currency value.

It took me a while to figure out how to interpret these currency values. Each cost is associated with a single use, and the duration of a single use increases as we move down the list of appliances. Since the designer assumes a fixed cost of electicity (shown in the footnote as 34p per kWh), at first, it seems like the costs should just increase from top to bottom. That's not the case, though.

Something else is driving these numbers behind the scene, namely, the intensity of energy use by appliance. The wifi router listed at the bottom is turned on 24 hours a day, and the daily cost of running it is just 6p. Meanwhile, running the fridge and freezer the whole day costs 41p. Thus, the fridge&freezer consumes electricity at a rate that is almost 7 times higher than the router.

The chart uses a split axis, which artificially reduces the gap between 8 hours and 24 hours. Here is another look at the bottom of the chart:



Let's examine the choice of "single use" as a common basis for comparing appliances. Consider this:

  • Continuous appliances (wifi router, refrigerator, etc.) are denoted as 24 hours, so a daily time window is also implied
  • Repeated-use appliances (e.g. coffee maker, kettle) may be run multiple times a day
  • Infrequent use appliances may be used less than once a day

I prefer standardizing to a "per day" metric. If I use the microwave three times a day, the daily cost is 3 x 3p = 9 p, which is more than I'd spend on the wifi router, run 24 hours. On the other hand, I use the washing machine once a week, so the frequency is 1/7, and the effective daily cost is 1/7 x 36 p = 5p, notably lower than using the microwave.

The choice of metric has key implications on the appearance of the chart. The bubble size encodes the relative energy costs. The biggest bubbles are in the heating category, which is no surprise. The next largest bubbles are tumble dryer, dishwasher, and electric oven. These are generally not used every day so the "per day" calculation would push them lower in rank.


Another noteworthy feature of the Bloomberg chart is the split legend. The colors divide appliances into five groups based on usage category (e.g. cleaning, food, utility). Instead of the usual color legend printed on a corner or side of the chart, the designer spreads the category labels around the chart. Each label is shown the first time a specific usage category appears on the chart. There is a presumption that the reader scans from top to bottom, which is probably true on average.

I like this arrangement as it delivers information to the reader when it's needed.




Modern design meets dataviz

This chart was submitted via Twitter (thanks John G.).


Perhaps the designer is inspired by this:


That's the Royal Ontario Museum, one of the beautiful landmarks in Toronto.


The chart addresses an interesting question - how much do home buyers over or under-estimate home value?  That said, gathering data to answer this question is challenging. I won't delve into this issue in this post.

Let's ask where readers are looking for data on the chart. It appears that we should use the right edge of each triangle. While the left edge of the red triangle might be useful, the left edges of the other triangles definitely would not contain data.

Note that, like modern architecture, the designer is playing with edges. None of the four right edges is properly vertical - none of the lines cuts the horizontal axis at a right angle. So the data actually reside in the imaginary vertical lines from the apexes to the horizontal baseline.

Where is the horizontal baseline? It's not where it is drawn either. The last number in the series is a negative number and so the real baseline is in the middle of the plot area, where the 0% value is.

The following chart shows (left side) the misleading signals sent to readers and (right side) the proper way to consume the data.


The degree of distortion is quite extreme. Only the fourth value is somewhat accurate, albeit by accident.

The design does not merely perturb the chart; it causes a severe adverse reaction.


P.S. [9/19/2022] Added submitter name.




Here's a radar chart that works, sort of

In the same Reuters article that featured the speedometer chart which I discussed in this blog post (link), the author also deployed a small multiples of radar charts.

These radar charts are supposed to illustrate the article's theme that "European countries are racing to fill natural gas storage sites ahead of winter."

Here's the aggregate chart that shows all countries:


In general, I am not a fan of radar charts. When I first looked at this chart, I also disliked it. But keep reading because I eventually decided that this usage is an exception. One just needs to figure out how to read it.

One reason why I dislike radar charts is that they always come with a lot of non-data-ink baggage. We notice that the months of the year are plotted in a circle starting at the top. They marked off the start of the war on Feb 24, 2022 in red. Then, they place the dotted circle, which represents the 80% target gas storage amount.

The trick is to avoid interpreting the areas, or the shapes of the blue and gray patches. I know, they look cool and grab our attention but in the context of conveying data, they are meaningless.

Redo_reuters_eugasradarall_1Instead of areas, focus on the boundaries of those patches. Don't follow one boundary around the circle. Pick a point in time, corresponding to a line between the center of the circle and the outermost circle, and look at the gap between the two lines. In the diagram shown right, I marked off the two relevant points on the day of the start of the war.

From this, we observe that across Europe, the gas storage was far less than the 80% target (recently set).

By comparing two other points (the blue and gray boundaries), we see that during February, Redo_reuters_eugasradarall_2gas storage is at a seasonal low, and in 2022, it is on the low side of the 5-year average. 

However, the visual does not match well with the theme of the article! While the gap between the blue and gray boundaries decreased since the start of the war, the blue boundary does not exceed the historical average, and does not get close to 80% until August, a month in which gas storage reaches 80% in a typical year.

This is example of a chart in which there is a misalignment between the Q and the V corners of the Trifecta Checkup (link).


The question/message is that Europeans are reacting to the war by increasing their gas storage beyond normal. The visual actually says that they are increasing the gas storage as per normal.


As I noted before, when read in a particular way, these radar charts serve their purpose, which is more than can be said for most radar charts.

The designer made several wise choices:

Instead of drawing one ring for each year of data, the designer averaged the past 5 years and turned that into one single ring (patch). You can imagine what this radar chart would look like if the prior data were not averaged: hoola hoop mania!


Simplifying the data in this way also makes the small multiples work. The designer uses the aggregate chart as a legend/how to read this. And in a further section below, the designer plots individual countries, without the non-data-ink baggage:


Thanks againto longtime reader Antonio R. who submitted this chart.

Happy Labor Day weekend for those in the U.S.!




Another reminder that aggregate trends hide information

The last time I looked at the U.S. employment situation, it was during the pandemic. The data revealed the deep flaws of the so-called "not in labor force" classification. This classification is used to dehumanize unemployed people who are declared "not in labor force," in which case they are neither employed nor unemployed -- just not counted at all in the official unemployment (or employment) statistics.

The reason given for such a designation was that some people just have no interest in working, or even looking for a job. Now they are not merely discouraged - as there is a category of those people. In theory, these people haven't been looking for a job for so long that they are no longer visible to the bean counters at the Bureau of Labor Statistics.

What happened when the pandemic precipitated a shutdown in many major cities across America? The number of "not in labor force" shot up instantly, literally within a few weeks. That makes a mockery of the reason for such a designation. See this post for more.


The data we saw last time was up to April, 2020. That's more than two years old.

So I have updated the charts to show what has happened in the last couple of years.

Here is the overall picture.


In this new version, I centered the chart at the 1990 data. The chart features two key drivers of the headline unemployment rate - the proportion of people designated "invisible", and the proportion of those who are considered "employed" who are "part-time" workers.

The last two recessions have caused structural changes to the labor market. From 1990 to late 2000s, which included the dot-com bust, these two metrics circulated within a small area of the chart. The Great Recession of late 2000s led to a huge jump in the proportion called "invisible". It also pushed the proportion of part-timers to all0time highs. The proportion of part-timers has fallen although it is hard to interpret from this chart alone - because if the newly invisible were previously part-time employed, then the same cause can be responsible for either trend.

_numbersense_bookcoverReaders of Numbersense (link) might be reminded of a trick used by school deans to pump up their US News rankings. Some schools accept lots of transfer students. This subpopulation is invisible to the US News statisticians since they do not factor into the rankings. The recent scandal at Columbia University also involves reclassifying students (see this post).

Zooming in on the last two years. It appears that the pandemic-related unemployment situation has reversed.


Let's split the data by gender.

American men have been stuck in a negative spiral since the 1990s. With each recession, a higher proportion of men are designated BLS invisibles.


In the grid system set up in this scatter plot, the top right corner is the worse of all worlds - the work force has shrunken and there are more part-timers among those counted as employed. The U.S. men are not exiting this quadrant any time soon.

What about the women?


If we compare 1990 with 2022, the story is not bad. The female work force is gradually reaching the same scale as in 1990 while the proportion of part-time workers have declined.

However, celebrating the above is to ignore the tremendous gains American women made in the 1990s and 2000s. In 1990, only 58% of women are considered part of the work force - the other 42% are not working but they are not counted as unemployed. By 2000, the female work force has expanded to include about 60% with similar proportions counted as part-time employed as in 1990. That's great news.

The Great Recession of the late 2000s changed that picture. Just like men, many women became invisible to BLS. The invisible proportion reached 44% in 2015 and have not returned to anywhere near the 2000 level. Fewer women are counted as part-time employed; as I said above, it's hard to tell whether this is because the women exiting the work force previously worked part-time.


The color of the dots in all charts are determined by the headline unemployment number. Blue represents low unemployment. During the 1990-2022 period, there are three moments in which unemployment is reported as 4 percent or lower. These charts are intended to show that an aggregate statistic hides a lot of information. The three times at which unemployment rate reached historic lows represent three very different situations, if one were to consider the sizes of the work force and the number of part-time workers.


P.S. [8-15-2022] Some more background about the visualization can be found in prior posts on the blog: here is the introduction, and here's one that breaks it down by race. Chapter 6 of Numbersense (link) gets into the details of how unemployment rate is computed, and the implications of the choices BLS made.

P.S. [8-16-2022] Corrected the axis title on the charts (see comment below). Also, added source of data label.

Four numbers, not as easy as it seems

Longtime reader Aleksander B. wasn't convinced by the following chart shown at the bottom of AFP's infographic about gun control.


He said:

Finally I was able to figure who got some support from NRA. But as a non-US citizen it was hard to get why 86% of republican tag points to huge red part. Then I figured out that smaller value of alpha channel codes the rest of republicans. I think this could be presented in some better way (pie charts are bad in presenting percentages of some subparts of the same pie chart - but adding a tag for 86% while skipping the tag for remaining 14% is cruel).

It's an example of how a simple chart with just four numbers is so hard to understand.


Here is a different view of the same data, using a similar structure as the form I chose for this recent chart on Swedish trade balance (link).


Funnels and scatters

I took a peek at some of the work submitted by Ray Vella's students in his NYU dataviz class recently.

The following chart by Hosanah Bryan caught my eye:

Rich Get Richer_Hosanah Bryan (v2)

The data concern the GDP gap between rich and poor regions in various countries. In some countries, especially in the U.K., the gap is gigantic. In other countries, like Spain and Sweden, the gap is much smaller.

The above chart uses a funnel metaphor to organize the data, although the funnel does not add more meaning (not that it has to). Between that, the color scheme and the placement of text, it's visually clean and pleasant to look at.

The data being plotted are messy. They are not actual currency values of GDP. Each number is an index, and represents the relative level of the GDP gap in a given year and country. The gap being shown by the colored bars are differences in these indices 15 years apart. (The students were given this dataset to work with.)

So the chart is very hard to understand if one focuses on the underlying data. Nevertheless, the same visual form can hold other datasets which are less complicated.

One can nitpick about the slight misrepresentation of the values due to the slanted edges on both sides of the bars. This is yet another instance of the tradeoff between beauty and precision.


The next chart by Liz Delessert engages my mind for a different reason.

The Rich Get Richerv2

The scatter plot sets up four quadrants. The top right is "everyone gets richer". The top left, where most of the dots lie, is where "the rich get richer, the poor get poorer".  This chart shows a thoughtfulness about organizing the data, and the story-telling.

The grid setup cues readers toward a particular way of looking at the data.

But power comes with responsibility. Such scatter plots are particularly susceptible to the choice of data, in this case, countries. It is tempting to conclude that there are no countries in which everyone gets poorer. But that statement more likely tells us more about which countries were chosen than the real story.

I like to see the chart applied to other data transformations that are easier. For example, we can start with the % change in GDP computed separately for rich and for poor. Then we can form a ratio of these two percent changes.



Metaphors give and take

Another submission came in from Euro Twitter. The following chart is probably from Germany:


As JB noted, this chart explains a financial pyramid scheme. I believe the numbers on the left are participants while the numbers on the right are the potential ill-gotten gains per person. The longer the pyramid scheme lasts, the more people participate, the more money flows to the top.

The pyramid is a natural metaphor for visualizing pyramid schemes. The levels of the pyramid correspond to levels of a pyramid scheme - the newly recruited participants expand the base while passing revenues up the pyramid.


The chart fails because it's not really a dataviz. There are exactly three bars that are scaled according to data. Everything else is presented as data labels.

Let's look at the two data series separately:


Each series is exponentially growing (in opposite directions). [Some of the data labels for participants may be incorrect.]

Unfortunately, the triangle is not a good medium to display exponential growth. In fact, the triangular structure imposes a linear growth constraint. The length of the base is directly proportional to the height from the top. As one traverses downwards level by level, the width of the base grows linearly - not exponentially.

To illustrate exponential growth, the edge of the triangle cannot be a straight line - it has to be s steep curve!


While natural, the pyramid metaphor is also severely restricting. The choice of chart form has unexpected consequences.


Who trades with Sweden

It's great that the UN is publishing dataviz but it can do better than this effort:


Certain problems are obvious. The country names turned sideways. The meaningless use of color. The inexplicable sequencing of the country/region.

Some problems are subtler. "Area, nes" - upon research - is a custom term used by UN Trade Statistics, meaning "not elsewhere specified".

The gridlines are debatable. Their function is to help readers figure out the data values if they care. The design omitted the top and bottom gridlines, which makes it hard to judge the values for USA (dark blue), Netherlands (orange), and Germany (gray).

See here, where I added the top gridline.


Now, we can see this value is around 3.6, just over the halfway point between gridlines.


A central feature of trading statistics is "balance". The following chart makes it clear that the positive numbers outweigh the negative numbers in the above chart.


At the time I made the chart, I wasn't sure how to interpret the gap of 1.3%. Looking at the chart again, I think it's saying Sweden has a trade surplus equal to that amount.