The Changing Colours of Rivers

No two rivers are the same, though they certainly can be similar. Rivers have their own ecosystems and when I was at school, I learned of the different classifications of rivers by the colour of their water: black, white, and clear. Broadly speaking, that just means the amount of sediment dissolved in the river’s water. Black colours appear when slow moving water has absorbed lots from its environment, think swamps. White waters resemble tea or coffee with added milk or cream. This happens when sediments enter and dissolved into the water. Clear water is that, relatively clear and free of sediment.

But a team of scientists at University of North Carolina at Chapel Hill (UNC Chapel Hill) recently released some work where they used shifts in blue to yellow and green to help classify rivers. Their classification differs, but broadly can point to a change from healthy (blue) to unhealthy (yellow and green). The novelty in their work, however, focuses on using satellite imagery to capture the colour of rivers and their evolution since the mid 1980s.

A look at the broader lower-48 of the United States

They published their findings as an interactive application driven primarily by a clickable map. Clearly not all rivers are available, but a large number are, and you can see some obvious patterns at a national scale—their work excludes Alaska and Hawaii. If blue represents healthy rivers, we see healthy rivers in New England and the Pacific Northwest with a host of green rivers in the Mid-Atlantic and Upper Midwest with yellow in the Mississippi basin and southeast.

I wanted to look at Pennsylvania a bit more specifically given my familiarity with the Commonwealth and zoomed in a bit on the map.

The colour of Pennsylvania’s rivers

You can see that using that above scale, Pennsylvania’s rivers are in okay, not great state. Some of the upper stretches of the Delaware and Susquehanna Rivers are coloured blue, but we mostly see a lot of green.

To the right of the map, the designers placed three smaller charts driven by the user’s selection of river. Let’s take a look at the Juniata River as an example—my grandfather grew up living alongside a tributary that emptied into the Frankstown Branch just a short walk from his house.

A look at the Schuylkill River south of the Fairmount Water Works

We can see that the chart on the upper right shows the colour shift over the decades for that observed section of the river. The legend provides the information that the section of the river has shifting blue—gotten healthier—and then below it looks for any seasonal changes. Here the chart is grey, indicating the system lacks enough data for a clear trend. This examines the short changes we might see in a river based on seasonal effects like rainy season, dry season, and human-driven effects—perhaps we pollute more in the spring and then use rivers recreationally in the summer.

Finally a distribution of the river section’s colour, all in wavelengths of light.

My biggest critique here would be the wavelengths. Users likely will not the colour spectrum by wavelength, and adding some labels like blue, yellow, and green could go a long ways to help users understand at what they are looking.

Overall, though, this is a really fascinating project.

Credit for the piece goes to John Gardner et al.

The Earth Is a Bit Bumpy

Last Friday I shared an xkcd post about the relative smoothness of the Earth. This week he posted an illustration but a slightly different scale. You can see more of Earth’s jagged edges.

Gotta love the Star Trek reference. I’m betting he used the length of the Kelvin timeline Enterprise, which I personally dislike, as it’s significantly larger than the prime timeline Enterprise of Shatner and Nimoy.

Anyway, Happy Friday, all.

Credit for the piece goes to Randall Munroe.

The Earth Is Actually Quite Smooth

At scale. Not quite as smooth as a billiards ball, as is often claimed. But still, with the majority of the Earth’s surface covered by water, the highest mountains of Everest and K2 make for mere fractions of differences in height relative to the Earth’s size.

But that did not stop xkcd from making a scale model of Earth.

Credit for the piece goes to Randall Munroe.

A Map of Unequal Comparisons

I’ve largely been busy creating and posting content on the Covid pandemic and its impact on the Pennsylvania, New Jersey, and Delaware tristate area along with, by request, both Virginia, and Illinois, my former home. It leaves me very little time for blogging, and I really do not want this site to become a blog of my personal work. That’s why I have a portfolio or my data project sites, after all.

But in posting my Covid datagraphics, I’ve come across variations of this map with all sorts of meme-y, witty captions saying why Canada is doing so much better than the US, why Americans shouldn’t be allowed to travel to Canada, and now why the Blue Jays shouldn’t be allowed to host Major League Baseball games.

Wait just a minute, there…

Well, that map isn’t necessarily wrong, but it’s incredibly misleading.

First, the map comes from the fantastic Johns Hopkins work on Covid-19. (Full disclosure, that’s the data source I use at work to create my work work datagraphics: https://philadelphiafed.org/covid-19/covid-19-research/covid-19-cases-and-deaths#.) And their site has a larger and more comprehensive dashboard (still hate that term but it does have sticking power) of which the map is the focal point.

The numbers as of this posting.

You can see the map there in the centre and some tables to the left, some tables to the right, and even a micro table beneath thundering away at the map’s position. I could get into the overall design—maybe I will one of these days—but again, let’s look at that map.

The crux of the argument is that there are a lot of red dots in the United States and very few in Canada. But look at the table in the dashboard on the left. At the very bottom you see three small tabs, Admin 0, Admin 1, and Admin 2. Admin 0 contains all entities at the sovereign state level, e.g. US, Canada, Sweden, Brazil, &c. Admin 1 is the provincial/state level, e.g. Pennsylvania, Illinois, Ontario, Quebec, &c. Admin 2 is the sub-provincial/sub-state level, e.g. Philadelphia County, Cook County, Chester County, Lake County, &c.

Notice anything about my examples? Not all countries have provinces/states, but Canada certainly does. And then at Admin 2, the examples and indeed the data only have US counties and US data. Everything in Canada has been aggregated up to Admin 1. And that is the problem.

The second part to point out is the dot-ness of the map. And to be fair, this is part of a broader problem I have been seeing in data visualisation the last few months. Dots, circles, or markers imply specificity in location. The centre of that object, after all, has to fall on a specific geographic place, a latitude and longitude coordinate. It utterly fails to capture the dimensions and physical size of the geographic unit, which can be critical.

Because not all geographic units are of the same size. We all know Rhode Island as one of the smallest US states. Let’s compare that to Nunavut or Yukon in Canada, massive provinces that spread across the Canadian Arctic. Rhode Island, according to Google, 1212 square kilometres. Nunavut? 808,200.

So now show both states/provinces on a map with one dot and Rhode Island’s will practically cover the state. And it will also be surrounded by and in close proximity to the states or Massachusetts and Connecticut. Nunavut, on the other hand will be a small dot in a massive empty space on a map. But those dots are equal.

Now, combine that with the fact that the Hopkins map is showing data on the US county level. Every single county in the United States gets a red dot. By default, that means the US is covered with red dots. But there is no county-level equivalent data for Canada. Or for Mexico (also seen in the above graphic). And so given we’re only using dots to relate the data, we see wide swaths of empty space, untouched by red dots. And that’s just not true.

Yes, large parts of the Canadian Arctic are devoid of people, but not southern Ontario and Quebec, not the southwestern coast of British Columbia, not the Maritimes.

The Hopkins map should be showing geographic units at the same admin level. By that I mean that when on Admin 0, the map should reflect geographic units of sovereign state level, allowing us to compare the US to Canada directly. But, and for this argument I’m assuming we’re keeping the dots despite their flaws, we only see Admin 0 level data.

Admin 1 shows only provincial level data. Some countries will begin to disappear, because Hopkins does not have the data at that level. But in North America, we still can compare Pennsylvania and Illinois to Ontario and Quebec.

But then at Admin 2, we only see the numerous dots of the United States counties. It’s neither an accurate nor a helpful comparison to contrast Chester County or Will County to the entire province of Ontario and so the map should not allow it. Instead, as the above graphic shows, it creates misconceptions of the true state of the pandemic in the US and Canada.

Credit for the Hopkins dashboard goes to, well, Hopkins.

Different Paths to Density

Yesterday we looked at the expansion of city footprints by sprawl, in modern years largely thanks to the automobile. Today, I want to go back to another article I’ve been saving for a wee bit. This one comes from the Economist, though it dates only back to the beginning of October.

This article looks at the different ways a city can achieve density. Usually one things of soaring skyscrapers, but there are other paths. For those interested, the article is a short read and I won’t cover it here. But for the sake of the graphic below, there are three basic paths: coverage, height, and crowding. Or to put in other terms, how much of the city is covered by homes, how tall those homes go, and how many people fit into each home.

Reticulating splines
Reticulating splines

I really like this graphic. It does a great job of using small multiples to compare and contrast three cities that exemplify the different paths. Notably, it keeps each city footprint at the same scale, making it easier to see things such as why Hong Kong builds skyward. Because it has little land. (It is, after all, an island and the tip of a peninsula.)

One area where I wish the graphic had kept to the small multiples is its display of Minneapolis. There, the scale shifts (note the lines for 5 km below vs. Minneapolis’ 10 km). I think I understand why, because the sprawling city would not have fit within the confines of the graphic, but that would have also hammered home the point of sprawl.

I should also point out that the article begins with a graphic I chose not to screenshot, but that I also really enjoy. It uses small multiples to compare cities density over time, running population on the x-axis and people per hectare on the y-. It is not a perfect graphic (it uses I think unnecessary arrowheads at the end of the line), but scatter plots over time are, I think, an underused graphic to show how two variables (ideally related) have moved in tandem over time.

Overall, this is a strong piece from the Economist.

Credit for the piece goes to the Economist graphics department.

Wicked Hot Islands

Though the temperatures might not always feel it, at least in Philadelphia, summer is ending and autumn beginning. Consequently I wanted to share this neat little work that explores urban heat islands. Specifically, this post’s author looks at Massachusetts and starts with a screenshot of the Boston area.

Wicked hot
Wicked hot

The author points out that the Boston Common and Public Garden are two areas of cool in an otherwise hot Boston. He also points out the Charles River and the divide between Boston and Brookline. I would like to add to it and point out the Fens and the Emerald Necklace.

I wonder if a scale of sorts would help, though the shift from warm yellows and reds to cooler greens and blues certainly helps differentiate between the cooler and warmer areas.

Credit for the piece goes to Krishna Karra.

Merging of the States

Dorian now speeds away from Newfoundland and into the North Atlantic. We looked at its historic intensity last week. But during that week, with all the talk of maps and Alabama, I noted to myself a map from the BBC that showed the forecast path.

Did New Jersey eat Delaware?
Did New Jersey eat Delaware?

But note the state borders. New Jersey and Delaware have merged. Is it Delawarsey? And what about Maryland, Virginia, and the District of Columbia? Compare that to this map from the Guardian.

Here the states are intact
Here the states are intact

What we have are intact states. But, and it might be difficult to see at this scale, the problem may be that it appears the BBC map is using sea borders. I wonder if the Delaware Bay, which isn’t a land border, is a reason for the lack of a boundary between the two states. Similarly, is the Potomac River and its estuary the reason for a lack of a border between Virginia, Maryland, and DC?

I appreciate that land shape boundary files are easy, but they sometimes can mislead users as to actual land borders.

Credit for these pieces go the BBC graphics department and the Guardian graphics department.

Urban Boom Towns

Today we look at a piece from the Guardian about the blossoming of some cities from, essentially, out of nowhere. Think similar to how there is really no reason for Las Vegas or Phoenix to exist—cities of hundreds of thousands situated smack in the middle of the desert. But most of these new growth cities, cities from scratch as the Guardian calls them, are sprouting in Africa and Asia.

The piece uses two pretty straight-forward graphics to show the scale of the growth problem.

A lot of urban area growth is yet to come.
A lot of urban area growth is yet to come.

I don’t love the area chart, but even for all its flaws, it it still massively obvious just how much Africa will contribute to population growth in the coming decades. And the line chart, which I find far more effective despite its borderline spaghetti-ness, shows just how much of that growth will likely be urban in nature.

But the star of the piece, for which you will need to click over to the original article to enjoy, are the motion graphics. They capture year-by-year the satellite views showing how the cities have grown from almost nothing. This is a screencapture of Ordos, China. But go back a couple of years and it’s almost an empty desert.

Check this out from decades ago and you'll see nothing.
Check this out from decades ago and you’ll see nothing.

Credit for the piece goes to Antonio Voce and Nick Van Mead.

The Entire United States Pt 2

Yesterday I wrote about the failure in a Politico piece to include Alaska and Hawaii in a graphic depicting the “entire” United States. After I had posted it, I recalled an article I read in the Guardian that looked at the shape of the United States, using the term “logo map”. It compared what many would consider the logo map to the actual map of the United States.

Still no New Zealand…
Still no New Zealand…

I warn you, it is a long read. But it was worth it to try and reframe the idea of what does the United States look like?

Credit for the piece goes to the Guardian graphics department.

The Entire United States

Last month Politico published an article called the Democrats’ Dilemma. It looked at what will likely be the crux of their debate for their 2020 candidates. Go moderate or hard left? The super simple version of the argument is that do you win by persuading independents and moderate Republicans to vote Democratic? Or do you win by ginning up the fervour of your liberal base and drive out the vote?

The article contrasts those approaches by looking at two neighbouring congressional districts. The first was won by Ilhan Omar, a Somali-American woman who has been at the centre of several causes célèbres in recent months. The second was won by a moderate, wealthy white man who has not really attracted any attention whatsoever.

But I don’t want to talk about the merits of either representative nor the fascinating split the article discusses. Instead, I want to look at a little piece of the graphics used in the article. It uses some simple stacked bar charts to compare and contrast the demographics of the representatives’ districts. Notably, they are different. But it goes on to compare and contrast them to the overall United States.

But what about New Zealand?
But what about New Zealand?

The first thing, I probably would have angled Mr. Phillips’ head so his head is straight, but that is a minor detail. The other thing I immediately noticed is a big pet peeve of mine. For the “Entire United States”, we have a map of the United States. Or do we?

What is missing? The entire states of Alaska and Hawaii, that’s what. I can understand not including Puerto Rico or other insular territories like the U.S. Virgin Islands because they are either not states or so small they would not appear visible at such a scale. However, Alaska and Hawaii are both integral parts of the United States. They are not marginal, like former Attorney General Jeff Sessions’ infamous quip about Hawaii being “some island in the Pacific”.

Perhaps at the above scale, Hawaii would be too small to appear—though I doubt it. But what about Alaska? It is the largest state. And Texas isn’t even a close second. So why is Alaska not included? Unfortunately—though fortunately for Politico, whose work I generally like—this is not a problem specific to Politico.

Even my own employer, the Federal Reserve Bank of Philadelphia, gets it wrong. One of their interactive data visualisation pieces, which for the record my team had nothing to do with, also completely omits Alaska and Hawaii in their map of the United States. And it’s a far larger map with ample space.

Still no New Zealand…
Still no New Zealand…

Including Alaska and Hawaii should not be afterthoughts. They are not second-class states. They are full constituent parts of the union. And if it is not easy to include them because they are not contiguous nor sharing the same continent, that should not obviate designers from including them in the United States.

Credit for the piece goes to the Politico’s design department and the Philadelphia Fed’s design department.