Last weekend was Easter—for both the Catholics and the Orthodox—and I visited the Appalachian ancestral home of the Carpatho–Rusyn side of my family. Before leaving town I drove up to the old cemetery on a hill overlooking the old church and the Juniata River to pay my respects to those who came before me and without whom I would not be here.
At the end of the four-hour drive back to Philadelphia, stuck in traffic on the Schuylkill Expressway because of course, I realised I had never really looked holistically at the causes of death of my direct ancestors. Earlier this week I spent some time putting that together and then, of course, I realised I wanted to see if I could find any patterns in the data. So of course I made a chart.
If we go back a couple of generations, you can see my ancestors lived to a median age of their mid-60s. But by the time of my grandparents that has increased to almost 80. Of course, the sample size is far smaller for grandparents than great-great-great-&c.-grandparents. Nonetheless, the general trend of the median line is upward.
A few exceptions pull those lines in both directions, however. Catherine Sexton died at the age of 35 from heart disease and James Scollon in the same generation died at 36 from typhoid fever. Additionally, that generation includes a few ancestors who remained in present-day Slovakia in what was one of the most impoverished areas of Europe. Not surprisingly they died in their 40s and 50s. If I exclude those people, the average shoots back up to about 70.
I also decided to colour the minimums and maximums by gender, because as you can see there is a broad pattern of longer-lived women and men who died young. I want to dig more into that aspect of the demographics at a later date to see if that trend holds. I suspect it would because that is the historical trend, but you never know.
After all the years of writing and publishing here on Coffeespoons, content centred on the sinking of RMS Titanic remains the most popular. And it was in the early hours of 15 April 1912 when she slipped beneath the surface of the North Atlantic. 700 people survived. 1500 people did not. Titanic’s sinking was the worst peacetime maritime disaster at the time and today, 113 years later, it remains likely in the top-five.
Each year I attempt to read something about the sinking. This year I’ve opted to re-read the first-hand account of Jack Thayer, a Philadelphian who survived the sinking. His father was an executive for the Pennsylvania Railroad and could afford first-class accommodations aboard Titanic.
I highly encourage you to read Thayer’s account of the sinking. First, it’s a short read in this era of short attention spans. But secondly, it provides the rare vantage point of a survivor who entered the water. With the temperature of the North Atlantic below freezing, almost without exception, everyone who entered the water died and so almost every passenger account is from someone who evacuated in a lifeboat.
In the aftermath of the sinking, Thayer managed to cling to a lifeboat and was eventually saved. When Thayer arrived in New York, a young art teacher named Lewis Skidmore worked with Thayer to illustrate what Thayer saw of Titanic on her final night.
Controversially, Thayer’s description included Titanic splitting in two. At the time, survivors intensely debated this point. Some claimed she went down intact and others, like Thayer, insisted she split in two. It was not until 1985 when she was discovered by Robert Ballard that the world learned Titanic did indeed split in two that fateful night, proving Thayer correct.
Unfortunately that was 40 years after Thayer’s death. He struggled with depression in the years that followed the sinking as other tragedies befell the man, including losing two of his own sons—one during World War II. He committed suicide at the age of 50 in 1945.
Skidmore’s graphics—black and white illustrations—arguably do as fine a job today as any complex 3D modelling could. They offer further proof that complex graphics need not necessarily communicate more clearly than sizzle and flash.
Last week I looked at baseball’s new torpedo bats in a humourous light. But I did want to take a more serious look, because bat evolution has been part of the tale of baseball since its beginning. Back in the day bats featured long lengths and heavier weights. These days, bats are in the mid-30 inch length and mid-30 ounce weight. Current regulations limit bats to 42 inches in length and a maximum diameter of 2.61 inches. 1 (There is some other stuff in there that is not terribly relevant to the torpedo bat.) And that’s it. Nothing about where the widest part of the bat need be nor the overall shape—merely a round, solid piece of wood.
The solid piece of wood is connected to corked bats. At my age I remember seeing the ESPN clips of Sammy Sosa’s corked bat in 2003 and the story of Albert Belle’s bat, stolen from the umpire’s room.
The torpedo bat, however, is not corked, instead it uses increased mass at the bat’s sweet spot, where it generates the greatest exit velocities when the batter hits the pitch. This graphic from ESPN shows how this works.
How a torpedo bat differs from the normal bat
Overall I like the graphic. The use of contrasting red and blue does a good job highlighting, particularly at the end where the normal bat continues with its edges running parallel as a regular cylinder. The torpedo bat changes into a cone with its top sliced off then rounded. I might have exaggerated the vertical dimensions of the bat illustration, but it still works.
Additionally, as I understand the design, it maximises the diameter of the bat to the aforementioned 2.61 inches. I have heard—but not confirmed—most bats do not reach the full diameter. If that is true, perhaps an illustration where the red lines fell below the maximum diameter of the torpedo bat could do a better job differentiating between the two shapes.
I prefer the above illustration to that produced by the Athletic/New York Times, which attempts a similar distinction.
New York Times version of the graphic.
I think ESPN’s overlay better shows the difference and that the Athletic’s wood pattern distracts from the graphic overall. Whereas ESPN uses the solid red vs. blue hatching to distinguish between the two shapes.
Of course that begets the question, why doesn’t every batter use the torpedo bat?
Suffice it to say, some players are better at hitting the ball consistently at the same spot on the bat. If a hitter can repeatedly make contact at a specific spot along the length of the bat, it makes sense to concentrate the mass of the bat at that spot for better hard contact. If, however, a hitter spreads his contact out along the length of the barrel, he probably wants a more evenly distributed mass to help create a better spread of good contact.
Regardless, as I wrote on Friday, the torpedo does not look like a torpedo. If anything, the normal baseball bat looks more like a torpedo than a torpedo bat.
Credit for the ESPN piece goes to ESPN’s graphics department.
Credit for the Athletic’s piece goes to Drew Jordan.
Baseball is back and so bring forth the controversies. One of the ones from last week? The torpedo bat. To be clear, the torpedo bat has been around for a few years—it’s not new. However, on Opening Weekend, the Yankees beat up on Milwaukee Brewers’ pitching. But a Yankees announcer specifically pointed out the design and the hype and the controversy was on.
But since this is a Friday, we’re going to look at a semi-humourous take from Davy Andrews of Fangraphs, a statistically-inclined baseball site. He uses illustrations to focus on the shape of the torpedo bat, which to my mind instantly did not look anything like a torpedo. (Read the full article for a funny aside about the shape’s resemblance—or lack thereof—to a torpedo.)
To be fair, I did not immediately think either old timey bomb or pregnant whale, but rather a bowling pin.
Yesterday I created a map detailing the new tariff rates released by President Trump on Wednesday. I was inspired by the curious inclusion of several small territories with almost no trade with the United States, and a few of whom are uninhabited. What follows is the graphic and the accompanying text I wrote as I wrote it.
I say that only because some people have not entirely caught the…let’s say tone with which I wrote.
All hail the new tariffs. Very obviously, foreign governments will be paying us lots of cash money. Places like Lesotho, with its so-called high rates of poverty, AIDS, and under-development, are clearly just fronts for the rich. Because their tariffs on us are turning them into the richest, most luxurious places on Earth.
Now I don’t know for sure, but some people say the shithole places like Nambia are really cash cows. Nerds tell me places like Nambia don’t exist, but their just idiots looking in the wrong wardrobe. Genius-level intellects like me can easily find Nambia on a map.
There are some very bad ombres out there, and I’m looking at you, Señor Diego Garcia. Some say you’re a thug with bad tattoos whom we should disappear to a secret black site. But the nerds keep telling me you’re not a person, just an island. That you’re not an illegal alien, but a British island where no civilians live, just US soldiers on a secret military base. But we need that money to pay for all the tax cuts for the rich. So we’ll just make our troops there pay Señor Garcia’s tariffs until he stops being lazy and pays us.
Then I’m looking at places like Christmas Island. That Santa Claus is really a bad guy. I know some of you like him—I like him too; he was good to me when I was a child. But all he does is export toys and joys. And that needs to be taxed. So I need Christmas Island to give us all their very real Christmas money.
Finally, I’m looking at Heard Island and McDonald Islands who’re trying to hide near the Antarctic Circle with all the other bad guys and their fortresses of solitude and vaults of swimmable coins. Sure, those nerds keep telling me these islands are uninhabited. But Amber Heard and Ronald McDonald are real people, in league with the Hamburgler, stealing all our rightful American money. The nerds say the islands are only inhabited by penguins. So if you want to say that Amber and Ronald are really just penguins, then we’re going to get all our sweet tariff money from the so-called penguins. Some of whom are emperors. Can you believe that? Emperor penguins? Emperors are rich. So we need to liberate those penguin dollars from the penguin monarchy.
Last week was baseball’s opening day. And so on the socials I released my predictions for the season and then a look at the revolving door that has been the Red Sox and second base since 2017.
Back in 2017 we were in the 11th year of Dustin Pedroia being the Sox’ star second baseman. That summer, Manny Machado slid spikes up into second and ruined Pedroia’s knee. Pedroia had surgery and missed Opening Day 2018 then struggled to return. He played 105 games in 2017 then only three in 2018 and then six in 2019. And thus began the instability. Here’s a list of the Opening Day second baseman since 2017.
2018 Eduardo Nuñez
2019 Eduardo Nuñez
2020 José Peraza
2021 Kiké Hernández
2022 Trevor Story
2023 Christian Arroyo
2024 Enmanuel Valdez
2025 Kristian Campbell
And, again, by comparison…
2007 Dustin Pedroia
2008 Dustin Pedroia
2009 Dustin Pedroia
2010 Dustin Pedroia
2011 Dustin Pedroia
2012 Dustin Pedroia
2013 Dustin Pedroia
2014 Dustin Pedroia
2015 Dustin Pedroia
2016 Dustin Pedroia
2017 Dustin Pedroia
But not only is it a lack of stability, it is a lack of production. Wins Above Replacement (WAR) is a statistic that attempts to capture a player’s value relative to an “average” player or substitute. A below replacement level person is less than 0 WAR. A substitute is 0–2, a regular everyday players is 2–5, an All Star is 5–8, and an elite MVP level performance is 8+ WAR. And, spoiler, the Sox have not had a 5+ WAR second baseman since Pedroia’s final full season in 2016.
Suffice it to say, the Sox have long had a need for a long-term second baseman. The graphics I created were meant to be two Instagram images in the same post, and so the the axis labels and lines stretch across the artboards.
The graphic shows pretty clearly the turmoil at the keystone. The two outliers are Kiké Hernández in 2021 and Trevor Story in 2022. The latter is easily explained. Story was signed to be the backup plan in case shortstop Xander Bogaerts left after 2022. (Back in 2013 I made a graphic after a similar revolving door of shortstops in the eight years after the Red Sox traded Nomar Garciaparra. Then the question was, would a young rookie named Xander Bogaerts be the replacement for the beloved Nomah. Xander played 10 years for the Sox.)
Kiké, however, is a bit trickier to explain. WAR weights value by position. A second baseman is worth more than a leftfielder. But shortstops and centrefielders are worth more than second baseman. And Kiké played a lot more shortstop and centre than he did second base, which likely explains his 4.9 WAR that season.
And so now in 2025 we had yet another guy starting at second. His name? Kristian Campbell. I saw him a few times last year as he rocketed from A to AAA, the lowest to highest levels of minor league player development below the major league. I thought he looked good and so did the professionals, because he’s a consensus top-10 prospect in the sport.
Going into Monday’s matchup between Boston and Baltimore, Campbell is hitting 6 for 14 with one homer and two doubles, an on-base percentage of .500 and an OPS (on-base plus slugging, which weights extra base hits more heavily than singles) of 1.286. Spoiler: that’s very good.
Boston beat writers are reporting the Sox and Campbell’s agent are in talks for a long-term extension.
It looks like the Sox may have found their new long-term second baseman.
This week began with Saint Patrick’s Day, a day that here in the States celebrates Ireland and Irish heritage. And I have an abundance of that. As we saw in a post earlier this year about some new genetic ancestry results, Ireland accounts for approximately 2/3 of my ancestry. But as many of my readers know, actual records-based genealogy is one of my big hobbies and so for this Saint Patrick’s Day, I decided to create a few graphics to capture all my current research on my family’s Irish heritage.
In the current political climate wherein we hyperfixate on immigration, I started with my ancestors’ immigration to North America.
My graphic features a timeline marking when certain ancestors arrived, with the massive caveat I do not know when all my Irish ancestors arrived. I separate the ancestors into paternal and maternal lines. My maternal lines are only half Irish, and unfortunately most of them offer little in terms of early records or origins and so the bulk of the graphic lands on my paternal lines.
I did sort out that two–four lines began in Canada and included them with orange dots. (The one couple married in Ireland shortly before setting sail for Canada. The other two lines married in Canada.) I also added a grey bar representing the length of the Great Famine. I suspect a number of my ancestors arrived during the famine based on the fact they begin to appear in the records around 1850, but sadly none of those records state when they arrived specifically instead they just appear in the United States.
I also used filled vs. open dots to indicate whether or not I had primary source documents for arrivals. I.e., a passenger manifest, naturalisation papers, &c. that specifically details immigration information weighs more heavily as evidence than, say, a census record wherein a respondent can say he or she arrived in such a year. (Spoiler, census records are not infallible.)
The overall takeaway, most of my Irish immigrants, for whom I have information, arrived in the middle of the 19th century within a decade of the Great Famine.
The second graphic features even more difficult data to find. Whence did my ancestors come?
For those unfamiliar with Irish genealogy, finding the town or parish from which your ancestors hailed can be nigh impossible. To start, you need some kind of American-based record that gives you a clue as to where in Ireland to look—a county or city. From my experience, most records simply state places of birth as “Ireland”—not very helpful.
Then if you can get back to Ireland, the typical resource you might use in the United States, United Kingdom, and other countries is the census. And Ireland did record a census every ten years, beginning in 1821. Unfortunately 1861 and 1871 were destroyed shortly after the data was recorded. Then during World War I, the 1881 and 1891 censuses were pulped due to a paper shortage. Then in 1921, there was no census because of the whole Irish Civil War thing. Finally in 1922, during the Battle of Dublin in the whole Irish Civil War thing, the Public Records Office at the Four Courts, which held government records dating back hundreds of years as well as guns and ammunition, was blown up. And with the ammunition, so too was blown up the census records for 1821, 1831, 1841, and 1851. In short, genealogists only have access to census records for 1901 and 1911. (The 1926 census organised post-Civil War, does not become public until 2027.)
Then you have the whole unavailability of Catholic Church records, which is another long discussion about the conflict between Protestants and Catholics in Ireland. (Just a minor thing in Irish history.)
There are some civil public records available and they begin in the mid-19th century, which in many cases is just a bit too late for genealogical purposes.
Suffice it to say, Irish genealogy can be tricky and in 15 years of researching it myself, I have only been able to find the origins of 10 Irish immigrant ancestors. For context, to the best of my knowledge I have 18 Irish immigrant ancestors. Thus that map is very empty.
The second map of the United States and United Kingdom is more complete because more complete records. It maps the residences of my Irish and Irish-American ancestors. Initially I attempted to link all the towns and cities with arrows to show the migration patterns, alas it quickly became a mess at such a small scale. That remains a project for another day.
My Irish heritage is a thing of which I am proud, and I am glad to say my genealogy hobby has allowed me to explore it much more deeply and richly than a green-dyed pint would allow.
Earlier this week, a Portuguese-flagged cargo container ship collided with an American-flagged tanker just off the Humber estuary in Yorkshire, England. The American-flagged ship, the Stena Immaculate, carries aviation fuel for the US Air Force. The Solong, the Portuguese-flagged tanker, carries alcohol, which is far better than the toxic chemicals initially feared.
We still know very little about the circumstances of the collision other than the Solong, travelling at 16 knots, slammed into the port side of the Stena Immaculate, which was anchored offshore.
I decided to write a little post because I enjoyed this graphic from the BBC, which details why the Stena Immaculate has not yet sunk—and at the time of my writing is not believed to be in danger of—despite the large hold amidships.
The graphic uses a simple line illustration of a bulk carrier in both 3/4 and a frontal view. The first shows how vessels like the Stena Immaculate separate their cargo into distinct holds, often watertight, so that, should a collision occur, the damage will not flood the entire ship or affect the load of the cargo. For the latter, sloshing liquids, as one example, can alter the centre of gravity and negatively impact ship stability.
The second line drawing illustrates the value of a double-hulled vessel wherein the outer hull shields the inner hull from puncture and prevents massive flooding of interior spaces.
Of course on 11 March, we are a little over a month away from the anniversary of the sinking of RMS Titanic. (In)famously in that case the critical issue was the same idea of watertight compartments. She had enough of them, but crucially they did not rise to the top of the ship as they would have necessarily impacted the luxury of first and second class accommodations. Titanic also did not have a double hull—her bottom was, but this did not run up the ship’s sides to the level where the iceberg impacted the ship.
Overall, I really like this graphic. It needs no elaborate and detailed illustration. Nor does it need sophisticated animations. All it uses is simple line illustrations.
Credit for the piece goes to the BBC graphics department.
Late last week I received an update on my ethnic breakdown from My Heritage, a competitor of Ancestry.com and other genealogy/family history/genetic ancestry companies. For many years, the genealogical community had been waiting for this long-promised update. And it has finally arrived.
For my money, My Heritage’s older analysis, v0.95, did not align with my historical record research—something I have done for almost 15 years now. That DNA analysis painted me with an 85% heritage of Irish, Scottish, and Welsh. Because I have spent a decade and a half researching my ancestors, I know all of my second-great-grandparents, 16 total. 85% means 13–14 of them would be Irish, Scottish, or Welsh. However, four of them are Carpatho-Rusyns from present day eastern Slovakia. And nowhere in my research have I found any connection to the Baltic states or Finland.
Compare that to the update.
Here we have a drastically reduced Irish component that, importantly, has been split from Scottish and Welsh, which now exists as its own genetic group. The East European group appears too low, but perhaps My Heritage identified some of my Slavic ancestry as Balkan—there is a sizeable Carpatho-Rusyn community in Vojvodina, an autonomous oblast in Serbia. Maybe Germanic too? That would start to push it near to 20%.
I do have English ancestry—my Angophilia must come from somewhere—though it is relatively small and I can trace it to the Medieval period. That includes more of the Norman elite than the Anglo-Saxon plebs and so seeing Breton register could be indicative of that Norman/Anglo-Saxon population mixture.
But how does My Heritage results compare to those provided by Ancestry and FamilyTreeDNA, two competitors whose services I have also used. And how does it compare to my actual historical document research?
My Heritage’s newest analysis certainly hits a lot better and is nearer to Ancestry, which aligns best with my research. I do have two questions for my second-great-grandparents. One surrounds Nathaniel Miller, one of whose grandparents (Eliza Garrotson) may not be English but rather Dutch from the Dutch colonisation of the Hudson River Valley in New York south of Albany.
The other question revolves around William Doyle. His mother is identified in the records variously as English and Irish. A family story on that side of the family also suggests one ancestor of English descent. And finally, a recently discovered marriage record for his parents details how his mother (Martha Atkins) was baptised and converted to Catholicism as an adult prior to her marriage. Not all Irish are Catholic, but the vast majority are and that would also suggest Martha was not Irish.
Taking those two questions into account, I have a small range of expected values for my English ancestry and a slightly larger one for my Irish and you can see those in the graphic.
When you compare that to the My Heritage results alongside the Ancestry and FamilyTreeDNA results you can see Ancestry aligns best with my research whereas FamilyTreeDNA aligns the least. My Heritage now falls squarely between the two. And so I consider their update a success. I think the company still has some work to do, but progress is progress.
Apologies, all, for the lengthy delay in posting. I decided to take some time away from work-related things for a few months around the holidays and try to enjoy, well, the holidays. Moving forward, I intend to at least start posting about once per week. After all, the state of information design these days provides me a lot of potential critiques.
Let us start with the news du jour , the application of tariffs on China and the delayed imposition on both Canada and Mexico. Firstly, let us be very clear what a tariff is. A tariff is a tax paid by importers or consumers on goods sourced from outside the country. In this case, we are talking about Canadian, Mexican, and Chinese imports and the United States slapping tariffs on goods from those countries. Foreign governments do not pay money to the United States, neither Canada, nor Mexico, nor China will pay money to the United States.
You will.
You should expect your shopping costs to increase, whether that is on the price of gasoline (imported from Canada), fast fashion apparel (from China), or avocados (from Mexico). On the more durable goods side, homes are built with Canadian lumber and your automobiles with parts sourced from across North America—the reason why we negotiated NAFTA back in the 1990s.
Now that we have established what tariffs are, why is the Trump administration imposing them? Ostensibly because border security and fentanyl. What those two issues have to do with trade policy and economics…I have no idea. But a few news outlets created graphics showing US imports from our top-five trading partners.
First I saw this graphic from the New York Times. It is a variation of a streamgraph and it needs some work.
A streamgraph type chart from the New York Times
To start, at any point along the timeline, can you roughly get a sense of what the value for any country is? No. Because there is no y-axis to provide a sense of scale. Perhaps these are the top import sources and these are their share of the total imports? Read the fine print and…no. These are the countries with a minimum of 2% share in 2024, which is approximately 75% of US imports.
This graphic fails at clearly communicating the share of imports. You need to somehow extrapolate from the y-height in 2024 given the three direct labels for Canada, Mexico, and China what the values are at any other point in time or for any other country.
Nevertheless, the chart does a few things nicely. It does highlight the three countries of importance to the story, using colours instead of greys. That focuses your attention on the story, whilst leaving other countries of importance still available for your review. Secondly, the nature of this chart ranks the greatest share as opposed to a straight stacked area chart.
Overall, for me the chart fails on a number of fronts. You could argue it looks pretty, though.
The aforementioned stacked area charts—also not a favourite of mine for this sort of comparison—forces the designer to choose a starting country in this case and then stack other countries atop it.
A stacked area chart from the BBC
What this chart does really well, especially well compared to the previous New York Times example is provide content for all countries across all time periods by the inclusion of the y-axis. Like the Times graphic it focuses attention on Canada, Mexico, and China with colour and uses grey to de-emphasise the other countries. You can see here how the Times’ decision to exclude all countries below 2% can skew the visual impact of the chart, though here all countries below Japan (everything but the top-five) are grouped as other.
Personally, the inclusion of the specific data labels for Canada, Mexico, and China distract from the visualisation and are redundant. The y-axis provides the necessary framework to visually estimate the share. If the reader needs a value to the precision level of tenths, a table may be a better option.
I could not find one of the charts I thought I had bookmarked and so in an image search I found a chart from one of my former employers on the same topic (though it uses value instead of share) and it is worth a quick critique.
A stacked area chart from Euromonitor International
Towards the end of my time there, I was creating templates for more wide-screen content. My fear from an information design and data visualisation standpoint, however, was the increased stretch in simple, low data-intensity graphics. This chart incorporates just 42 data points and yet it stretches across 1200 pixels on my screen with a height of 500.
Compare that to the previous BBC graphic, which is also 1200 pixels, but has a greater height of 825 pixels. Those two dimensions give ratios of 2.4 for Euromonitor International and 1.455 for the BBC. Neither is the naturally aesthetically pleasing golden ratio of 1.618, but at least the BBC version is close to Tufte’s recommended 1.5–1.6. The idea behind this is that the greater the ratio, the softer the slope of the line. This can make it more difficult to compare lines. A steeper slope can emphasise changes over time, especially in a line chart. You can roughly compare this by looking at the last few years of the longer time span in the BBC graphic to the entirety of this graphic. You can more easily see the change in the y-axis because you have more pixels in which to show the change.
Finally we get to another New York Times graphic. This one, however, is a more traditional line chart.
A line chart from the New York Times
And for my money, this is the best. The data is presented most clearly and the chart is the most legible and digestible. The colours clearly focus your attention on Canada, Mexico, and China. The use of lines instead of stacked area allow the top importer to “rise” to the top. You can track the rapid rise of Chinese imports from the late 1990s through to the first Trump administration and the imposition of tariffs in 2018—note the significant drop in the line. In fact you can see the impact in Mexico becoming the United States’ top trading partner in recent years.
Over the years, if I had a dollar for every time I was told someone wanted a graphic made “sexier” or with more “sizzle” or made “flashier”, I would have…a bigger bank account. The issue is that “cooler” graphics do not always lead to clearer graphics. Graphics that communicate the data better. And the guiding principle of information design and data visualisation should be to make your graphics clear rather than cool.
Credit for the New York Times streamgraph goes to Karl Russell.
Credit for the BBC graphic goes to the BBC graphics department.
Credit for the Euromonitor International graphic goes to Justinas Liuima.
Credit for the New York Times line chart goes to the New York Times.
