Metaphors, maps, and communicating data

There are some data visualization that are obviously bad. But what makes them bad?

Here is an example of such an effort:

This visualization of carbon emissions is not successful. There is precious little that a reader can learn from this chart without expensing a lot of effort. It's relatively easy to identify the largest emitters of carbon but since the data are not expressed per-capita, the chart mainly informs us which countries have the largest populations. 

The color of the bubbles informs readers which countries belong to which parts of the world. However, it distorts the location of countries within regions, and regions relative to regions, as the primary constraint is fitting the bubbles inside the shape of a foot.

The visualization gives a very rough estimate of the relative sizes of total emissions. The circles not being perfect circles don't help. 

It's relatively easy to list the top emitters in each region but it's hard to list the top 10 emitters in the world (try!) 

The small emitters stole all of the attention as they account for most of the labels - and they engender a huge web of guiding lines - an unsightly nuisance.

The diagram clings dearly to the "carbon footprint" metaphor. Does this metaphor help readers consume the emissions data? Conversely, does it slow them down?

A more conventional design uses a cartogram, a type of map in which the positioning of countries are roughly preserved while the geographical areas are coded to the data. Here's how it looks:

I can't seem to source this effort. If any reader can find the original source, please comment below.

This cartogram is a rearrangement of the footprint illustration. The map construct eliminates the need to include a color legend which just tells people which country is in which continent. The details of smaller countries are pushed to the bottom. 

In the footprint visualization, I'd even consider getting rid of the legend completely. This means trusting that readers know South Africa is part of Africa, and China is part of Asia.

Imagine: what if this chart comes without a color legend? Do we really need it?

***

I'd like to try a word cloud visual for this dataset. Something that looks like this (obviously with the right data encoding):

(This map is by Michael Tompsett who sells it here.)

 

Pies, bars and self-sufficiency

Andy Cotgreave asked Twitter followers to pick between pie charts and bar charts:

The underlying data are proportions of people who say they won't get the coronavirus vaccine.

I noticed two somewhat unusual features: the use of pies to show single proportions, and the aspect ratio of the bars (taller than typical). Which version is easier to understand?

To answer this question, I like to apply a self-sufficiency test. This test is used to determine whether the readers are using the visual elements of the chart to udnerstand the data, or are they bypassing the visual elements and just reading the data labels? So, let's remove the printed data from the chart and take another look:

For me, these charts are comparable. Each is moderately hard to read. That's because the percentages fall into a narrow range at one end of the range. For both charts, many readers are likely to be looking for the data labels.

Here's a sketch of a design that is self-sufficient.

The data do not appear on this chart.

***

My first reaction to Andy's tweet turned out to be a misreading of the charts. I thought he was disaggregating the pie chart, like we can unstack a stacked bar chart.

Looking at the data more carefully, I realize that the "proportions" are not part to the whole. Or rather, the whole isn't "all races" or "all education levels". The whole is all respondents of a particular type.

 

 

Making graphics last over time

Yesterday, I analyzed the data visualization by the White House showing the progress of U.S. Covid-19 vaccinations. Here is the chart.

John who tweeted this at me, saying "please get a better data viz".

I'm happy to work with them or the CDC on better dataviz. Here's an example of what I do.

Obviously, I'm using made-up data here and this is a sketch. I want to design a chart that can be updated continuously, as data accumulate. That's one of the shortcomings of that bubble format they used.

In earlier months, the chart can be clipped to just the lower left corner.

Reading an infographic about our climate crisis

Let's explore an infographic by SCMP, which draws attention to the alarming temperature recorded at Verkhoyansk in Russia on June 20, 2020. The original work was on the back page of the printed newspaper, referred to in this tweet.

This view of the globe brings out the two key pieces of evidence presented in the infographic: the rise in temperature in unexpected places, and the shrinkage of the Arctic ice.

A notable design decision is to omit the color scale. On inspection, the scale is present - it was sewn into the graphic.

I applaud this decision as it does not take the reader's eyes away from the graphic. Some information is lost as the scale isn't presented in full details but I doubt many readers need those details.

A key takeaway is that the temperature in Verkhoyansk, which is on the edge of the Arctic Circle, was the same as in New Delhi in India on that day. We can see how the red was encroaching upon the Arctic Circle.

***

Next, the rapid shrinkage of the Arctic ice is presented in two ways. First, a series of maps.

The annotations are pared to the minimum. The presentation is simple enough such that we can visually judge that the amount of ice cover has roughly halved from 1980 to 2009.

A numerical measure of the drop is provided on the side.

Then, a line chart reinforces this message.

The line chart emphasizes change over time while the series of maps reveals change over space.

This chart suggests that the year 2020 may break the record for the smallest ice cover since 1980. The maps of Australia and India provide context to interpret the size of the Arctic ice cover.

I'd suggest reversing the pink and black colors so as to refer back to the blue and pink lines in the globe above.

***

The final chart shows the average temperature worldwide and in the Arctic, relative to a reference period (1981-2000).

This one is tough. It looks like an area chart but it should be read as a line chart. The darker line is the anomaly of Arctic average temperature while the lighter line is the anomaly of the global average temperature. The two series are synced except for a brief period around 1940. Since 2000, the temperatures have been dramatically rising above that of the reference period.

If this is a stacked area chart, then we'd interpret the two data series as summable, with the sum of the data series signifying something interesting. For example, the market shares of different web browsers sum to the total size of the market.

But the chart above should not be read as a stacked area chart because the outside envelope isn't the sum of the two anomalies. The problem is revealed if we try to articulate what the color shades mean.

On the far right, it seems like the dark shade is paired with the lighter line and represents global positive anomalies while the lighter shade shows Arctic's anomalies in excess of global. This interpretation only works if the Arctic line always sits above the global line. This pattern is broken in the late 1990s.

Around 1999, the Arctic's anomaly is negative while the global anomaly is positive. Here, the global anomaly gets the lighter shade while the Arctic one is blue.

One possible fix is to encode the size of the anomaly into the color of the line. The further away from zero, the darker the red/blue color.

 

 

Book Review: Visualizing with Text by Richard Brath

The creative process is sometimes described in terms of diverge-converge cycles. The diverge step involves experimentation and rewards suspending disbelief, while excesses are curbed and concepts refined during the converge step. Richard Brath's just-released book Visualizing with Text is an important resource that expands our appreciation for the place of text in visual displays.

Books on data visualization fall into recognizable types, of which two popular ones are the style guide, such as Edward Tufte, Dona Wong, and Alberto Cairo, and the coding manual, such as Ben Fry (processing) and Hadley Wickham (ggplot, Shiny). Brath's volume belongs to neither of those - it reads more like an encyclopedic catalog of how text can be incorporated into charts and graphs. He challenges us to blow up our imaginative space for characters, words, sentences, paragraphs and prose. It is a valuable aid for the diverge step of our creative process.

In modern data visualization, text is treated as an accessory, frequently found in titles, labels, legends, footnotes or surrounding text. Brath wants us to elevate text to the starring attraction. Starting with baby steps, such as direct labeling of lines and objects, and coordinating colors between chart elements and words, he experiments with inserting text into unlikely crannies, not shying away from ideas that even he admits may be somewhat of a dead-end.

One of the more immediately useful examples is the use of text labels that hug the lines on a line chart, similar to how roads and rivers are labeled on maps. I wish all software developers implement this function without delay.

A more esoteric example is to replace these lines with small-size text, as Brath makes an analogy between sentences and lines.

I am still deciding if this is a gold mine or a minefield. It is thought-provoking nonetheless.

Finally, the book includes some flights of fancy, like this one:

The red superscripts are numeric codes for French departments (provinces), arranged in ascending order of a given metric, and placed in proportional distance within the prose!

The converge step is left to the reader, as Brath refrains from bullhorning his opinions about chart types, which is why readers should not expect a style guide. He includes many experimental graphics, and may provide the pros and cons of a form without registering a judgement.

Because many of these ideas have yet to enter the mainstream, we'd need to implement these ideas on our own, which is why readers will not find a coding manual. As mentioned above, even the simplest and least controversial tactic of directly labeling lines is not available in Excel, let alone text that hugs or replaces lines. (This proves Brath's point that our community has done text a disservice.) Other ideas explored in later chapters require such features as italicizing numeric proportions of a word, rather than the entire word.

Recently, text has become a mainstay of Big Data. Visualizing with Text is timely, relevant and provocative. It is also clearly written, and tightly organized. Chapter 13 neatly summarizes the key concepts that have appeared along the way. There are plenty of use cases, primarily derived from research or business. After reading this book, you'll revel in the new sandbox of text, and long to free yourself from the constraints of your tool.

***

I recommend that you get the paper copy of the book. I reviewed the electronic version, and what irony! As you may have guessed, the electronic version ruins the typesetting. On every page, certain paragraphs show up in tiny font that resist all attempts to magnify, making Brath's case that legibility is an important metric for text visualization. Some of the more unusual fonts are dropped. The images are too small, even when popped up.

[P.S. Richard has a webpage where he included larger images and some code.]

Election visual 3: a strange, mash-up visualization

Continuing our review of FiveThirtyEight's election forecasting model visualization (link), I now look at their headline data visualization. (The previous posts in this series are here, and here.)

It's a set of 22 maps, each showing one election scenario, with one candidate winning. What chart form is this?

Small multiples may come to mind. A small-multiples chart is a grid in which every component graphic has the same form - same chart type, same color scheme, same scale, etc. The only variation from graphic to graphic is the data. The data are typically varied along a dimension of interest, for example, age groups, geographic regions, years. The following small-multiples chart, which I praised in the past (link), shows liquor consumption across the world.

Each component graphic changes according to the data specific to a country. When we scan across the grid, we draw conclusions about country-to-country variations. As with convention, there are as many graphics as there are countries in the dataset. Sometimes, the designer includes only countries that are directly relevant to the chart's topic.

***

What is the variable FiveThirtyEight chose to vary from map to map? It's the scenario used in the election forecasting model.

This choice is unconventional. The 22 scenarios is a subset of the 40,000 scenarios from the simulation - we are left wondering how those 22 are chosen.

Returning to our question: what chart form is this?

Perhaps you're reminded of the dot plot from the previous post. On that dot plot, the designer summarized the results of 40,000 scenarios using 100 dots. Since Biden is the winner in 75 percent of all scenarios, the dot plot shows 75 blue dots (and 25 red).

The map is the new dot. The 75 blue dots become 16 blue maps (rounded down) while the 25 red dots become 6 red maps.

Is it a pictogram of maps? If we ignore the details on the maps, and focus on the counts of colors, then yes. It's just a bit challenging because of the hole in the middle, and the atypical number of maps.

As with the dot plot, the map details are a nice touch. It connects readers with the simulation model which can feel very abstract.

Oddly, if you're someone familiar with probabilities, this presentation is quite confusing.

With 40,000 scenarios reduced to 22 maps, each map should represent 1818 scenarios. On the dot plot, each dot should represent 400 scenarios. This follows the rule for creating pictograms. Each object in a pictogram - dot, map, figurine, etc. - should encode an equal amount of the data. For the 538 visualization, is it true that each of the six red maps represents 1818 scenarios? This may be the case but not likely.

Recall the dot plot where the most extreme red dot shows a scenario in which Trump wins 376 out of 538 electoral votes (margin = 214). Each dot should represent 400 scenarios. The visualization implies that there are 400 scenarios similar to the one on display. For the grid of maps, the following red map from the top left corner should, in theory, represent 1,818 similar scenarios. Could be, but I'm not sure.

Mathematically, each of the depicted scenario, including the blowout win above, occurs with 1/40,000 chance in the simulation. However, one expects few scenarios that look like the extreme scenario, and ample scenarios that look like the median scenario.  

So, the right way to read the 538 chart is to ignore the map details when reading the embedded pictogram, and then look at the small multiples of detailed maps bearing in mind that extreme scenarios are unique while median scenarios have many lookalikes.

(Come to think about it, the analogous situation in the liquor consumption chart is the relative population size of different countries. When comparing country to country, we tend to forget that the data apply to large numbers of people in populous countries, and small numbers in tiny countries.)

***

There's a small improvement that can be made to the detailed maps. As I compare one map to the next, I'm trying to pick out which states that have changed to change the vote margin. Conceptually, the number of states painted red should decrease as the winning margin decreases, and the states that shift colors should be the toss-up states.

So I'd draw the solid Republican (Democratic) states with a lighter shade, forming an easily identifiable bloc on all maps, while the toss-up states are shown with a heavier shade.

Here, I just added a darker shade to the states that disappear from the first red map to the second.

Twitter people UpSet with that Covid symptoms diagram

Been busy with an exciting project, which I might talk about one day. But I promised some people I'll follow up on Covid symptoms data visualization, so here it is.

After I posted about the Venn diagram used to depict self-reported Covid-19 symptoms by users of the Covid Symptom Tracker app (reported by Nature), Xan and a few others alerted me to Twitter discussion about alternative visualizations that people have made after they suffered the indignity of trying to parse the Venn diagram.

To avoid triggering post-trauma, for those want to view the Venn diagram, please click here.

[In the Twitter links below, you almost always have to scroll one message down - saving tweets, linking to tweets, etc. are all stuff I haven't fully figured out.]

Start with the Questions

Xan’s final comment is especially appropriate: "There's an over-riding Type-Q issue: count charts answer the wrong question".

As dataviz designers, we frequently get locked into the mindset of “what is the best way to present this dataset?” This line of thinking leads to overloaded graphics that attempt to answer every possible question that may arise from the data in one panoptic chart, akin to juggling 10 balls at once.

For complex datasets, it is often helpful to narrow down the list of questions, and provide a series of charts, each addressing one or two questions. I’ll come back to this point. I want to first show some of the nicer visuals that others have produced, which brings out the structure and complexity of this dataset.

 

The UpSet chart

The primary contender is the “UpSet” chart form, as best exemplified by Bart’s effort

The centerpiece of this chart is the matrix of dots. The horizontal rows of dots represent the presence of specific symptoms such as cough and anosmia (loss of smell and taste). The vertical columns are intuitive, once you get it. They represent combinations of symptoms, and the fill/no-fill of the dots indicates which symptoms are being combined. For example, the first column counts people reporting fatigue plus anosmia (but nothing else).

The UpSet chart clearly communicates the structure of the data. In many survey questions (including this one conducted by the Symptom Tracker app), respondents are allowed to check/tick more than one answer choices. This creates a situation where the number of answers (here, symptoms) per respondent can be zero up to the total number of answer choices.

So far, we have built a structure like we have drawn country outlines on a map. There is no data yet. The data are primarily found in the sidebar histograms (column/bar charts). Reading horizontally to the right side, one learns that the most frequently reported symptom was fatigue, covering 88 percent of the users.* Reading vertically, one learns that the top combination of symptoms was fatigue plus anosmia, covering 16 percent of users.

***

Now come the divisive acts.

Act 1: Bart orders the columns in a particular way that meets his subjective view of how he wants readers to see the data. The columns are sorted from the most frequent combinations to the least. The histogram has a “long tail”, with most of the combinations receiving a small proportion of the total. The top five combinations is where the bulk of the data is – I’d have liked to see all five columns labeled, without decimal places.

This is a choice on the part of the designer. Nils, for example, made two versions of his UpSet charts. The second version arranges the combinations from singles to quintuples.

 

Digression: The Visual in Data Visualization

The two rendering of “UpSet” charts, by Nils and Bart, is a perfect illustration of the Trifecta Checkup framework. Each corner of the Trifecta is an independent dimension, and yet all must sync. With the same data and the same question types, what differentiates the two versions is the visual design.

See how many differences you can find, and make your own design choices!

 

I place the digression here because Act 1 above has to do with the Q corner, and both visual designs can accommodate the sorting decisions. But Act 2 below pertains to the V corner.

Act 2: Bart applies a blue gradient to the matrix of dots that reinforces his subjective view about identifying frequent combinations of symptoms. Nils, by contrast, uses the matrix to show present/absent only.

I’m not sure about Act 2. I think the addition of the color gradient overloads the matrix in the chart. It has the nice effect of focusing the reader’s attention on the top 5 combinations but it also requires the reader to have understood the meaning of columns first. Perhaps applying the gradient to the histogram up top rather than the dots in the matrix can achieve the same goal with less confusion.

 

Getting Obtuse

For example, some readers (e.g. Robin) expressed confusion.

Robin is alleging something the chart doesn’t do. He pointed out (correctly) that while 16 percent experienced fatigue and anosmia only (without other symptoms), more than 50 percent reported fatigue and anosmia, plus other symptoms. That nugget of information is deeply buried inside Bart’s chart – it’s the sum of each column for which the first two dots are filled in. For example, the second column represents fatigue+anosmia+cough. So Robin wants to aggregate those up.

Robin’s critique arises from the Q(uestion) corner. If the designer wants to highlight specific combinations that occur most frequently in the data, then Bart’s encoding makes perfect sense. On the other hand, if the purpose is to highlight pairs of symptoms that occur most frequently together (disregarding symptoms outside each pair), then the data must be further aggregated. The switch in the Question requires more Data manipulation, which then affects the Visualization. That's the essence of the Trifecta Checkup framework.

Rest assured, the version that addresses Robin’s point will not give an easy answer to Bart’s question. In fact, Xan whipped up a bar chart in response:

This is actually hard to comprehend because Robin’s question is even hard to state. The first bar shows 87 percent of users reported fatigue as a symptom, the same number that appeared on Bart’s version on the right side. Then, the darkened section of the bar indicates the proportion of users who reported only fatigue and nothing else, which appears to be about 10 percent. So 1 out of 9 reported just fatigue while 8 out of 9 who reported fatigue also experienced other symptoms.

 

Xan’s bar chart can be flipped 90 degrees and replace Bart’s histogram on top of the matrix. But you see, we end up with the same problem as I mentioned up top. By jamming more insights from more questions onto the same chart, we risk dropping the other balls that were already in the air.

So, my advice is always to first winnow down the list of questions you want to address. And don’t be afraid of making a series of charts instead of one panoptic chart.

***

Act 3: Bart decides to leave out labels for the columns.

This is a curious choice given the key storyline we’ve been working with so far (the Top 5 combinations of symptoms). But notice how annoying this problem is. Combinations require long text, which must be written vertically or slanted on this design. Transposing could help but not really. It’s just a limitation of this chart form. For me, reading the filled dots underneath the columns as column labels isn’t a show-stopper.

 

Histograms vs Bar Charts

It’s worth pointing out that the sidebar “histograms” are not both histograms. I tend to think of histograms as a specific type of bar (column) chart, in which the sum of the bars (columns) can be interpreted as a whole. So all histograms are bar charts but only some bar charts are histograms.

The column chart up top is a histogram. The combinations of symptoms are disjoint, and the total of the combinations should be the total number of answer choices selected by all respondents. The bar chart on the right side however is not a histogram. Each percentage is a proportion to the whole, and adding those percentages yields way above 100%.

I like the annotation on Bart’s chart a lot. They are succinct and they give just the right information to explain how to read the chart.

 

Limitations

I already mentioned the vertical labeling issue for UpSet charts. Here are two other considerations for you.

The majority of the plotting area is dedicated to the matrix of dots. The matrix contains merely labels for data. They are like country boundaries on a map. While it lays out the structure of data very clearly, the designer should ask whether it is essential for the readers to see the entire landscape.

In real-world data, the “long tail” phenomenon we saw earlier is very common. With six featured symptoms, there are 2^6 = 64 possible combinations of symptoms (minus 1 if they filtered out those not reporting symptoms*), almost all of which will be empty. Should the low-frequency columns be removed? This is not as controversial as you think, because implicitly both Bart and Nils already dropped all empty combinations!

 

Data and Code

Kieran Healy left a comment on the last post, and you can find both the data (thank you!) and some R code for UpSet charts at his blog.

Also, Nils has a Shiny app on Github.

 

(*) One must be very careful about what “users” are being represented. They form a tiny subset of users of the Symptom Tracker app, just those who have previously taken a diagnostic test and have self-reported at least one symptom. I have separately commented on the analyses of this dataset by the team behind the app. The first post discusses their analytical methods, the second post examines how they pre-processed the data, and a future post will describe the data collection practices. For the purpose of this blog post, I’ll ignore any data issues.

(#) Bart’s chart is conceptual because some of the columns of dots are repeated, and there is one column without fills, which should have been removed by a pre-processing step applied by the research team.

Habit-busting designs don't work

The design changes that most frustrate users are those that bust their habits.

Case in point. Apple re-designed the bottom navigator of the iphone mail app. See what it looked like before and what it looks like today:

Notice how the 2nd slot from the bottom right used to be for replying, and after the re-design, it has become the button for deleting. So when I intended to reply to a message, my finger instinctively presses that 2nd button and now, instead of replying, the message gets deleted!

In the last few years, my finger hit that button thousands of times whenever the brain said to reply. Now, it's really hard to change this habit. I kept having to undo the delete. It's frustrating beyond belief.

This also shows the habit is in the muscle memory, and I'm no longer paying attention to the visual icon. A more direct dataviz analogy is when you belatedly discovered that the horizontal axis in a line chart isn't representing time because you didn't read the axis labels.

***

A similar thing happened inside an elevator (lift) recently.

Most elevator panels place the Door Open and Door Close buttons side by side. Typically, the Door Open is on the left and the Door Close is on the right.

This particular elevator panel has the Door Open button on top, and Door Close at the bottom, laid out vertically. To the right of the Door Open button is the Alarm button! So I sounded the Alarm when I intended the doors to close.

(I didn't take a photo at the time. The figure on the right is a rough sketch of what the panel looked like.)

I bet the alarm is pressed multiple times a day by mistake.

Bubble charts, ratios and proportionality

A recent article in the Wall Street Journal about a challenger to the dominant weedkiller, Roundup, contains a nice selection of graphics. (Dicamba is the up-and-comer.)

The change in usage of three brands of weedkillers is rendered as a small-multiples of choropleth maps. This graphic displays geographical and time changes simultaneously.

The staircase chart shows weeds have become resistant to Roundup over time. This is considered a weakness in the Roundup business.

***

In this post, my focus is on the chart at the bottom, which shows complaints about Dicamba by state in 2019. This is a bubble chart, with the bubbles sorted along the horizontal axis by the acreage of farmland by state.

Below left is a more standard version of such a chart, in which the bubbles are allowed to overlap. (I only included the bubbles that were labeled in the original chart).

The WSJ’s twist is to use the vertical spacing to avoid overlapping bubbles. The vertical axis serves a design perogative and does not encode data.  

I’m going to stick with the more traditional overlapping bubbles here – I’m getting to a different matter.

***

The question being addressed by this chart is: which states have the most serious Dicamba problem, as revealed by the frequency of complaints? The designer recognizes that the amount of farmland matters. One should expect the more acres, the more complaints.

Let's consider computing directly the number of complaints per million acres.

The resulting chart (shown below right) – while retaining the design – gives a wholly different feeling. Arkansas now owns the largest bubble even though it has the least acreage among the included states. The huge Illinois bubble is still large but is no longer a loner.

Now return to the original design for a moment (the chart on the left). In theory, this should work in the following manner: if complaints grow purely as a function of acreage, then the bubbles should grow proportionally from left to right. The trouble is that proportional areas are not as easily detected as proportional lengths.

The pair of charts below depict made-up data in which all states have 30 complaints for each million acres of farmland. It’s not intuitive that the bubbles on the left chart are growing proportionally.

Now if you look at the right chart, which shows the relative metric of complaints per million acres, it’s impossible not to notice that all bubbles are the same size.

Conceptualizing a chart using Trifecta: a practical example

In response to the reader who left a comment asking for ideas for improving the "marginal abatements chart" that was discussed here, I thought it might be helpful to lay out the process I go through when conceptualizing a chart. (Just a reminder, here is the chart we're dealing with.)

First, I'm very concerned about the long program names. I see their proper placement in a horizontal orientation as a hard constraint on the design. I'd reject every design that displays the text vertically, at an angle, or hides it behind some hover effect, or abbreviates or abridges the text.

Second, I strongly suggest re-thinking the "cost-effectiveness" metric on the vertical axis. Flipping the sign of this metric makes a return-on-investment-type metric, which is much more intuitive. Just to reiterate a prior point, it feels odd to be selecting more negative projects before more positive projects.

Third, I'd like to decide what metrics to place on the two axes. There are three main possibilities: a) benefits (that is, the average annual emissions abatement shown on the horizontal axis currently), b) costs, and c) some function that ties together costs and benefits (currently, this design uses cost per unit benefit, and calls it cost effectivness but there are a variety of similar metrics that can be defined).

For each of these metrics, there is a secondary choice. I can use the by-project value or the cumulative value. The cumulative value is dependent on a selection order, in this case, determined by the criterion of selecting from the most cost-effective program to the least (regardless of project size or any other criteria).

This is where I'd bring in the Trifecta Checkup framework (see here for a guide).

The decision of which metrics to use on the axes means I'm operating in the "D" corner. But this decision must be made with respect to the "Q" corner, thus the green arrow between the two. Which two metrics are the most relevant depends on what we want the chart to accomplish. That in turn depends on the audience and what specific question we are addressing for them.

Fourth, if the purpose of the chart is exploratory - that is to say, we use it to guide decision-makers in choosing a subset of programs, then I would want to introduce an element of interactivity. Imagine an interface that allows the user to move programs in and out of the chart, while the chart updates itself to compute the total costs and total benefits.

This last point ties together the entire Trifacta Checkup framework (link). The Question being exploratory in nature suggests a certain way of organizing and analyzing the Data as well as a Visual form that facilitates interacting with the information.