Grumble about the graphics choices: dark-grey-on-black-with-other-dark-colors is a terrible color scheme, that renders the borders nearly invisible.
There's a reason print maps have a standard set of colors, with very light blue for oceans, white for land backgrounds, and a variety of dark colors for features. The "modern white-on-black web aesthetic" only really works for text- and figure-heavy pages, where you must then use very light colors (white, yellow, light orange, light green) for features/lines.
Interestingly, this website reliably crashes my firefox on linux while consuming 55GB of memory.
Claude's TLDR of what's causing the problem (may or may not be accurate): "That animation loop is almost certainly leaking memory: each time-step it draws new border geometry (GeoJSON/vector shapes) but doesn't free the old frames, so RAM climbs without bound. When you interact — especially auto-playing the timeline — the tab grows until it swallows all 62 GB of RAM + swap and the kernel kills it."
or perhaps, some Plasma bug was fixed in Plasma 6.7x. Who knows...
EDIT: Now I see you installed it via flatpak. That is also a source of endless inconsistencies and instabilities. So you should maybe try the native fedora rpm.
Cool visuals, as with everything like this where the creator probably just churned open datasets through LLMs there are many inaccuracies particularly around borders.
An interesting effort though, and at least this one has a decent page about sourcing.
Yes, definitely. Claude and other models might produce minor differences in design outputs, but overall they apply similar principles. This has been harped on in many threads over the past several months.
I completely disagree. It does a really good job of showing the relative sizes of political entities throughout history and how those things impact the total amount of conflict. I might suggest that perhaps your politics is impacting how you view this data because perhaps it doesn't align with your ideology. It would be useful to overlay total population sizes only this data as well. That would bring out the historical trends much more accurately. And it is ignoring these political and technology trends that leads to the most wrong things humanities professors will try to claim. Perhaps that's why you don't like it. It disproves graphically a lot of ideological viewpoints.
PS Pinker talks a lot about these types of trends. Maybe look him up for better explanations.
I did. Yep. Tried to source everything to credit original creators and make this as transparent as possible.
Getting borders exactly right has been pretty crazy to wrangle. Just doing it with LLMs hasn't proven a good route so I've been hand-correcting things slowly but surely. Long road to go.
Very interesting and watchable. Do you differentiate between wars and "conflicts"? There's so many of the latter and everyone seems to avoid the term "war".
"Conflict" seemed like the broader useful definition. Especially in modern times, defining it as just "wars" felt like it'd leave too much out; and I really wanted to map conflict to show how impactful it's been on societies and cultures.
I guess the (war?) elephant in the room is that written history as something that attempts to record a somewht balanced, comprehensive account of an event is a modern, western, anomaly.
There are a lot of very old written histories recording various battles. For example, the Spring and Autumn Annals have a somewhat detailed account of the Battle of Chengpu and its aftermath: https://ctext.org/chun-qiu-zuo-zhuan/xi-gong#comm18160 This map actually briefly flashes a red dot at 632 BC, but since it's not part of any named war, you could easily miss it.
The areas where you see fewer wars don't necessarily lack written historical records, it might just be that nobody bothered to translate those records into a machine-readable format yet. (I'd guess this map is based on Wikidata.)
China is the one exception actually. India, Africa and the New World civilizations didn't really like focusing on the past and didn't record it. Europe and China did. In China that's probably about the strength of the bureaucracy. In Europe its probably more about kings establishing legitimacy.
Wikipedia has a known bias of having mostly the historical events present that western world has written down. This map seems almost balanced, how come it used a better or at least perhaps fairer datasource?
Sigh, because certain civilizations wrote things down more than others. Also, some environments don't lend themselves to preserving things, others are. These will naturally create biases because if you are studying somewhere wet, almost nothing not made of stone will survive, even and especially things that are written upon. In the desert, things will be preserved for centuries, including things made of clay, leather or paper.
This is VERY wrong. Almost all estimates go for at least 3x higher Russian casualties than
Ukrainian. Russia has been attacking for 4+ years just throwing bodies at the problem with Ukrainians defending with technology. Where do these estimates even come from? Makes me question the validity of the information on this site
That is the Western media narrative anyway. The casualty rate for the current mode of offensive warfare with small infantry teams infiltrating under cover of darkness (e.g. Ukraine for the latter part of the 2023 counteroffensive, Russia in its more recent offensive in the Donbas) has been extremely high on both sides. But I'm pretty sure Russia had a favorable kill ratio during their 2022 summer offensive in the Donbas where they just pummeled fortifications with standoff weapons like the Buratino, and many of Ukraine's most experienced troops died in that offensive.
Anyway, the Western stereotype of "Russian human wave attacks" is mostly wrong. Even when Russia is just throwing bodies into the fray (like the convict troops in Bakhmut), those can't really be described as "human wave" tactics (again, they're small infantry teams infiltrating at night). And Ukraine has thrown lots of hastily mobilized cannon fodder at the front as well: look for videos of protesting TDF soldiers and their relatives on Telegram if you don't believe me.
What are you talking about? Mediazone already has 230k confirmed! russian deaths with names. Current confirmed rate of ~30k/month casualities for Russia couldn't be sustainable for much smaller population of Ukraine if they were trading 1:1 or even worse
Numbers range from 323 k to 2 mio. total war deaths since 2022.
Hugh leap.
But the numbers for Ukraine and russian federation seem to be swapped: Ukraine does not even have 600 k soldiers, so probably in 10-20 or so years cannot have these many loses without front collapse.
You said human waves, not the person you are replying to. Russian did throw bodies at the problem and has shown callous disregard for the welfare of its soldiers.
I feel like you're not really replying to the comment above.
At the start of the war there was plenty of meat being thrown into the trenches on the Ukraine side. Have you not seen any videos of men being press-ganged off the streets or in their homes? Or of the protesting wives of TDF units thrown into the trenches with virtually no training and only token weapons and ammo? (You can find plenty of both on Ukrainian Telegram channels.)
If Ukraine is treating their soldiers better now, it’s at least partly because they’re running out of them.
Again, I feel like you're not really replying to the comment above. Or mine, really.
We are talking about the numbers being way off, and you are seemingly picking a bone with Ukraine. It’s like you are having an entirely separate and somewhat related discussion.
It is not human waves, but they have been on the attack for quite some time now relying on small infantry units while obtaining very little ground. What does that tell you about what's happening to those infantry?
Impressive that people still try to pretend there's any grand strategy here when it's easy to watch 100+ new Russians (mostly old dudes, not in a team, not at night) get FPV'd every single day now. Tell us about the 'Kyiv feint' next
The website cites "Wikipedia: Russo-Ukrainian War; Ukrainian Ministry data" as its source, but the figures in the overview (600,000+ Ukrainian military deaths, 100,000+ Russian) aren't close to anything in either source [1].
The UI looks like what Claude would generate, and I'd guess the data was scraped and compiled the same way - but it's not clear how it could reference a Wikipedia page and use numbers that contradict it.
Unless russian efforts to flood the internet with English-language news sites pushing the russian narrative into training sets and retrieval [2] are working. There is even a website tracking the known russian networks that flood the internet with news articles [3].
War only goes down when the latest one killed the existing armies too effectively, so that it takes a generation until the armies can be replenished and the memory of the realities of war forgotten. Disease like the plague can also cause a reduction in war numbers since people have other things to worry about than conquering or destroying their neighbors.
I recently figured that Spain went on a war for 700 years, just to carry on in the Arauco war for another 300 years, thus, literally being at war for 1000 years.
FF just crashed immediately. Twice. Which kinda surprised me in a good way, because usually when it struggles for memory it just hangs there for a couple of minutes until it gets killed by the OS. Cannot remember it being killed that easily by another website.
Love this. Acurate display of humanity and what it does when things run out. Would be cool if ithad a ethnicity map showing a cultures get transported elsewhere or vanish (e.g. the minority purges over the last hundred years in the middle east).
71 comments
[ 2.5 ms ] story [ 53.7 ms ] threadRobinson Projection would be much more accurate.
Imagine foreign policy being distorted by the Mercator projection.
There's a reason print maps have a standard set of colors, with very light blue for oceans, white for land backgrounds, and a variety of dark colors for features. The "modern white-on-black web aesthetic" only really works for text- and figure-heavy pages, where you must then use very light colors (white, yellow, light orange, light green) for features/lines.
Claude's TLDR of what's causing the problem (may or may not be accurate): "That animation loop is almost certainly leaking memory: each time-step it draws new border geometry (GeoJSON/vector shapes) but doesn't free the old frames, so RAM climbs without bound. When you interact — especially auto-playing the timeline — the tab grows until it swallows all 62 GB of RAM + swap and the kernel kills it."
Curiously, the website works just fine in chrome on android.
Blink monopoly strikes again, I guess.
Maybe it's a firefox bug that was introduced somewhere after 140? idk
EDIT: Now I see you installed it via flatpak. That is also a source of endless inconsistencies and instabilities. So you should maybe try the native fedora rpm.
the explosion of dots from the civil war is really something!
An interesting effort though, and at least this one has a decent page about sourcing.
PS Pinker talks a lot about these types of trends. Maybe look him up for better explanations.
Getting borders exactly right has been pretty crazy to wrangle. Just doing it with LLMs hasn't proven a good route so I've been hand-correcting things slowly but surely. Long road to go.
The areas where you see fewer wars don't necessarily lack written historical records, it might just be that nobody bothered to translate those records into a machine-readable format yet. (I'd guess this map is based on Wikidata.)
> Estimates: 600,000+ Ukrainian military deaths; 100,000+ Russian deaths; 30,000-40,000 civilian deaths.
This is VERY wrong. Almost all estimates go for at least 3x higher Russian casualties than Ukrainian. Russia has been attacking for 4+ years just throwing bodies at the problem with Ukrainians defending with technology. Where do these estimates even come from? Makes me question the validity of the information on this site
Anyway, the Western stereotype of "Russian human wave attacks" is mostly wrong. Even when Russia is just throwing bodies into the fray (like the convict troops in Bakhmut), those can't really be described as "human wave" tactics (again, they're small infantry teams infiltrating at night). And Ukraine has thrown lots of hastily mobilized cannon fodder at the front as well: look for videos of protesting TDF soldiers and their relatives on Telegram if you don't believe me.
But the numbers for Ukraine and russian federation seem to be swapped: Ukraine does not even have 600 k soldiers, so probably in 10-20 or so years cannot have these many loses without front collapse.
I feel like you're not really replying to the comment above.
If Ukraine is treating their soldiers better now, it’s at least partly because they’re running out of them.
We are talking about the numbers being way off, and you are seemingly picking a bone with Ukraine. It’s like you are having an entirely separate and somewhat related discussion.
The UI looks like what Claude would generate, and I'd guess the data was scraped and compiled the same way - but it's not clear how it could reference a Wikipedia page and use numbers that contradict it.
Unless russian efforts to flood the internet with English-language news sites pushing the russian narrative into training sets and retrieval [2] are working. There is even a website tracking the known russian networks that flood the internet with news articles [3].
[1] https://waratlas.org/#year=2026&conflict=russo-ukrainian-war...
[2] https://www.theguardian.com/world/2025/nov/21/english-langua...
[3] https://portal-kombat.com/
https://en.wikipedia.org/wiki/Toledo_War
But to be fair, this is really cool.
ublock origin on mobile isn't worth it when brave exists.