I don’t like snakes. I have a (seemingly) primordial bias against them.
My dislike for them only increased when, as a child I stepped on a Gardner snake barefoot while playing tag in my yard. Not poisonous, but unexpected and creepy. (I imagine the snake enjoyed the run-in even less.) I still shriek whenever I see a Garner snake on my property, which is quite frequently, as I moved into a pine forest full of them.
I’ve always been fascinated by poisonous snakes, though, ever since seeing the movie, Raiders of the Lost Ark, and because where I grew up, we didn’t have anything “venomous” to speak of. Not even spiders.
I think that’s why snakes feel “more dangerous” than they probably really are. Bears, moose, and sometimes a random wolf or two were frequently in my yard growing up… (albeit a spruce forest then). Our crab apple tree tended to attract momma bears with young cubs every year. Made waiting for the school bus intriguing. But they were no problem. They were mammals. Our species’ cousins, if you will.
But poisonous snakes?! I felt so lucky I never had to deal with those.
Today, I decided to create a map of poisonous snake species around the world to celebrate a day that (inwardly) freaks me out – World Snake Day. Perhaps it will help me overcome my fear. (And yes, I know snakes are crucial for ecosystems. I still don’t like them.)
Animal range datasets are frequently murky. Also, the datasets aren’t generally easily or readily available or accurate for 60 different species around the world at once.
So today’s map was a two-part challenge for me.
The first: use Spatial Data Doctor GPT to find, collect, and trace different toxic snake species ranges anywhere it could find them. Naturally, I asked it to also return reliability scores based on how each feature was retrieved, its source, and the level of generalization the source had to begin with.
Simple. This was a much bigger ask than many other data retrievals I’ve done, because in some cases I believe it had to analyze print maps from open-access journal articles, etc., and then logically determine and trace those ranges.
Well done, Spatial Data Doctor GPT. Well done. This is my favorite GPT in the whole WebMapGPT suite. I don’t think many GIS companies have realized this yet, but there is absolutely little reason for a human cartographer in 2026 to ever visit a slow-loading, archaic GIS data portal again. That’s what LLM agents are for! Try it if you haven’t yet. It may just change your dataset creation life. Don’t forget to ask for a source and reliability score for each feature returned too and/or a provenance file. Particularly because the anti-LLM trolls are out there, and they are looking for prompt cartographers to strike. Beware of the trolls. They’re like… like… what we’re mapping today!
So where was I… yes, poisonous snakes!
I now had an epic dataset. So it was time to create a map. Alas… you may have heard, but over the past two weeks I’ve been creating a lot of stuff for other prompt cartographers. What drives me recently is helping more people develop prompt cartography workflows and become more proficient at the process. That’s my entire modus operandi right now. That’s why I started creating map design styles and schemas a couple of weeks ago.
On that front, I’ve also created a Map Design Lab at www.promptcartography.com that allows you to mix and mash up to four of these design schemas to create your own Franken-Design schema (or style genome) using what I am currently calling a Map DNA composite score. (I labeled the tool “beta” on purpose to reserve the right to change the subpar naming conventions I tend to come up with at one in the morning on a Monday).
So now we all have over 40 researched styles we can use, as well as a site where you can mix and match different styles together, add new colors, remove some palette colors, edit and tweak which fonts are used, etc. So the second goal of today’s map…
I wanted to use some new schemas and a Map Design Lab mash-up to make this map. I will show you three mash-up schemas below. All of them are based off of board game systems I own and love, because I haven’t had time to play many games recently.
In the end, I kept the the World War II Undaunted Normandy-inspired card game theme mixed with the classic Cassini French Cartography style. I just preferred the legend styling and map organization with on this one compared to the other outputs. It’s a map style mash-up! Cool!
You’ll noticed that the board game schemas aren’t posted anywhere online yet. There’s a reason for that. They were made with a new WebMapGPT agent I’ve built called Map Styler. These were the first design schemas I created with the actual GPT. They are prototype and QA outputs. The original design schemas were designed in a different place, not using a specialized GPT. I took what I learned from making those to create a GPT tool I can share with the world.
I will release Map Styler shortly – maybe tomorrow depending on how fried I am after work. It will allow you to write a simple prompt and create detailed design schemas (essentially what are used in the Map Design Lab) of your own.
For example, in the examples shown below, I asked the GPT to create schemes based off board games I was looking at on a cart in my garage. You can make design schema from almost anything – the backs of DVD covers work, Ravesnburger puzzle box layouts, Ikea furniture catalog designs… Yup, I’ve been playing around a lot. You can make some seriously cool designs by mixing them together!
And rest assured, the tool never copies any trademarked or copyrighted icons, logos, etc. It creates a map style scheme inspired by what you ask it to do research on – not copying directly. It’s amazing. It’s fun. And it just sped up the cartographic process significantly! Anyone who has played Bohnanza knows that the overly saturated yellow map with eye-burning red font is spot on. Design inspiration and mixing is merely a prompt away now. No coding required. No GUIs required. Just a half-decent prompt.
Map Styler will be announced on the WebMapGPT Blog soon. So if you’re interested, stay tuned.
But anyway… snakes! Blasted snakes! Yikes!
Today’s map shows where 60 of the world’s most toxic snakes live (roughly).
It’s interactive and designed for kids, parents, and people who randomly might Google or LLM “show me a map of the world’s most poisonous snakes please.” I’m sure this will land on the 47th page of Google search, but it will be out there, in the wild.
Another prompt cartography artifact. Ta-da!
I hope you enjoy today’s map. It was fun to make. And I can’t wait to get to the next day, because snakes… yuck!
Bohnanza / Family Card-Inspired Game Mash Up Styled Map
Forest Shuffle Game / Cassini Map Mash Up
(Also inspired by Guillaume Touya and the very nice cheese-shop guy in Nantes 🙂
The Final Design Version
World War II / Undaunted-Card-Game-System-Inspired Mash Up Map
Pop-up style was designed in one-off prompt without me mentioning it at all, because the schema associated the card styles and layouts to map components.
Web Mapper Prompt (www.webmapgpt.com) Single Prompt for each style.
Please create a fun, educational map about snake ranges for World Snake Day. The audience should be targeted at school children who are fascinated by deadly snakes and may ask Google to find a map. (Add some SEO stuff to the metadata at top, please, about poisonous snake maps, ranges, interactive maps, reliable, etc..) I’m attaching a ton of information and data created by another GPT. Use the data that makes the most sense for the easiest, cleanest presentation on an apache website. There is a readme file that describes what the datasets are and what you can visualize. If there is data on number of deaths, etc., perhaps add proportional circles, etc. Otherwise, just ranges are fine. Perhaps take a look at the data and attached style schema before you begin and provide your best style and dsign suggestion. Thank you!