Growing a World Wide Web

Every year Telegeography publishes a comprehensive, annually updated map that visualizes active and planned submarine telecommunications cables around the world. The futuristic looking 2025 Submarine Cable Map was released back in January. This map shows the current extent of the world’s active submarine telecommunications cables and those currently under construction.

The Internet Infrastructure Map allows you to see how the current network of submarine telecommunication cables has developed over the last 36 years. The map combines two key elements of the internet: submarine fiber-optic cables, which connect continents across the ocean floor, and Internet Exchange Points (IXPs), the data hubs where networks interconnect and exchange traffic.

One of the standout features of the Internet Infrastructure Map is its animated timeline. Starting from 1989, when the first modern subsea cables appeared, you can step year by year through time and watch the growth of the global internet unfold on the map. The animation reveals how over time the sparse early connections developed into the dense intercontinental webs of cables of today, with new systems being added almost every year. By sliding through to 2025, you can literally watch the internet’s backbone expand, seeing when major routes were built and how new IXPs shifted regional connectivity patterns.

How Big is Anything?

The Size of Anything is an interactive map that lets you compare the sizes of different locations around the world.

To be honest, when I first heard about The Size of Anything I thought, “Not another size comparison map.” Just off the top of my head I can think of several similar tools:

• The True Size Of ... 
• GeoSize Compare 
• Country Size Comparison 
• Reprojector

- all of which let you directly compare the scale of different geographical areas on the same map.

However, as soon as I started playing with The Size of Anything, my mood quickly improved. It’s very well done. Perhaps the most impressive feature is the sheer number of locations you can compare. As long as OpenStreetMap knows about it, The Size of Anything can handle it. That means the options are enormous: parks, airports, islands, neighborhoods, towns, cities - almost anything mapped on OSM.

There’s also a fun extra: if you select the Treasure button, you’ll find a few non-geographic objects to overlay, such as a blue whale, the Titanic, or an airplane.

So, while The Size of Anything isn’t exactly a brand-new concept, it takes the idea and executes it brilliantly. With its huge range of locations and playful extras, it’s probably one of the best tools out there for exploring and comparing the true scale of different geographical areas.

Letters Found on the Moon

This isn’t a tale about Lunar correspondence, but of letters shaped from the craters and shadows of the Moon. Alphabet Moon uses imagery of Lunar contours and ridges to shape a typeface out of unfamiliar terrain. Each letter is drawn not with ink but with the valleys, peaks, and scars of the lunar surface, transforming geological history into the letters of the alphabet.

Enter your name - or any other word - into Alphabet Moon, and watch it spelled out in letters drawn from the Moon’s ancient terrain. Each character is carefully matched to the shape of a crater, ridge, or valley, so that what begins as a simple word is reimagined in a language etched into the lunar surface.

Beneath each lunar letter lies a short explanation of how that form was created. These notes not only reveal the exact location on the Moon where the feature can be found, but also describe the geological forces that shaped it - whether an ancient impact, the slow cooling of lava, or the shifting of the Moon’s crust.

Alphabet Moon is a brilliant reinterpretation of Rhett Dashwood’s Google Maps Typography. Back in 2009, Dashwood unveiled an “Earth font” made up of 26 satellite images of our planet, each one resembling a different letter of the alphabet.

NASA appreciated Dashwood’s idea so much that they went on to create their own interactive typewriter, allowing you to write your name using satellite imagery. Type a name into Your Name in Landsat and watch it spelled out in Earth features captured by Landsat satellites. You can even download an image of your word written in massive Earth letters, and by hovering over each letter you can discover where in the world those shapes occur.

And if letters alone don’t satisfy your curiosity, there’s also Earth Clock - an online digital clock that uses satellite images of natural features resembling numbers to display the current time wherever you are.

Changing Parking Lots to Homes

Over recent years, a number of urban planning maps have revealed just how much valuable city land is devoted to surface parking lots. For example, the Parking Lot Map highlights the percentage of land in U.S. city centers taken up by parking.

The School of Cities at the University of Toronto has gone a step further with its project From Parking Spaces to Living Spaces. Using a compelling story map, the school shows how Toronto could repurpose underutilized surface parking lots into new housing. This shift could help address the Toronto housing crisis, while also generating significant property tax revenue for the city.

The interactive map illustrates:

• How much land in Toronto is currently used for parking lots.
• How much of this space sits within 1 km of public transit stations.
• How much additional revenue could be created if these lots were converted into city-owned housing developments.

The report makes a clear case: Toronto’s surface parking lots, often located in prime, transit-friendly neighborhoods, produce little revenue while the city struggles with an acute shortage of affordable housing. Redeveloping these sites into well-designed residential communities would not only increase tax revenue but also create much-needed homes and build more vibrant, complete neighborhoods.

Using AI to Search Maps

The magnificent David Rumsey Map Collection now has an AI Search Assistant that can help you find maps, learn more about individual maps, and even query specific elements within maps.

The David Rumsey Map Collection is one of the largest online collections of maps, and its new AI Search Assistant is a fantastic resource. It not only helps you search and discover maps in the collection but also lets you dive deeper into individual maps and the cartographers who created them.

1. Find Maps in the Collection

From the David Rumsey Map Collection home page, you can now use the AI Search Assistant to find maps based on themes, locations, or any other criteria you can imagine. The Assistant will return a list of maps in the collection that match your search.

Examples of queries you might use are:

• Show me maps of ancient Rome
• Show panoramic maps of New York 
• Show maps that feature compass roses

Ask what projection is used in a map!

2. Ask Questions About an Individual Map

When viewing a specific map, click the blue chat icon in the bottom-right corner to open the AI Search Assistant. You can then ask technical or contextual questions about the map you’re looking at. 

For example:

• What map projection is used by this map?
• Can you translate the map title into English?
• Who was Mercator?

What is this building?

3. Ask Questions about Map Details

When viewing a map you can zoom in on details to ask questions about a specific section of the map. 

Examples of these types of queries might be:

• What was happening at the Colosseum in the time depicted?
• What is this building? 
• What do the words below the map say?

Tourist Minesweeper

Tourist Minesweeper is a twist on the classic puzzle game, using a gridded map of real locations to highlight the spread of Airbnb in popular Spanish tourist destinations. If you’ve ever played Minesweeper (and who hasn’t?), you’ll recognize the rules - but here, instead of dodging bombs, you’re “sweeping” for zones of tourist pressure.

Currently you can play Tourist Minesweeper on gridded maps of Mallorca and Barcelona (more locations are on the way). It is worth noting that the visualization rules are a little different in each location: 

• In Mallorca, a cell has a mine if it contains more than 15 Airbnb listings, turning clusters of vacation rentals into hotspots on the map. 
• The Barcelona map, on the other hand, uses price as its metric: any grid cell with an average nightly rate above €200 is treated as a mine. 

This shift between quantity in Mallorca and affordability in Barcelona highlights different facets of the tourism problem - the sheer density of rentals in one case, and the economic pressure of high prices in the other.

The grid cells in each city are also based on different sized areas. I believe that in Barcelona each grid cell represents a 500 m square, while in Mallorca each grid cell is a 2 km² area.

Via: Quantum of Sollazzo

Roll Your Own Geocoder

One of my favorite recent side-projects has been Meet Cute, a playful web map that generates tiny “micro-romance” stories whenever you click on a location. The conceit is simple: click on a map, and out comes a love story set in the nearest town.

But behind that simple experience was a not-so-romantic technical problem: finding the name of the nearest town.

At first, I leaned on the Overpass API, which is built on top of OpenStreetMap. Every time a user clicked somewhere on the map, my code would:

1. Send a query to Overpass.

2. Ask it to return the nearest place=city|town|village node.

3. Use the closest name to plug into my romance grammar.

This worked beautifully… sometimes.

Overpass is a shared resource, and as many OSM developers know, it can be slow when overloaded. On bad days, a single click could stall Meet Cute for several seconds. Worse, I didn’t want to keep hammering the Overpass servers every time someone got carried away clicking for story after story.

---

Building My Own Geocoder

So I decided to roll my own lightweight client-side geocoder.

I realised that instead of calling Overpass for every click, I could pre-build a list of towns and cities and serve it as a static file. In fact I knew that TripGeo has just such a list of over 11,000 towns and cities around the world, that it uses for its daily Scrambled Maps Challenge. 

Here’s what I did:

• I grabbed the TripGeo dataset of world cities (latitude, longitude, name).

• I saved it into a cities.json file and hosted it on GitHub Pages.

• I wrote a little Geocoder class that loads this JSON into memory and, given a latitude/longitude (defined by a user map click), finds the closest matching city by brute force distance calculation.

const geocoder = new CityGeocoder('https://mapsmania.github.io/geocoder/cities.json');
const town = geocoder.reverse(lat, lon);

The entire geocoder lives in a single JavaScript class which is loaded into Meet Cute and then called when the map needs to find a new location.

Now, when the user clicks on the map, my code doesn’t need to make a network call at all. It just looks up the nearest city locally. The result is that Meet Cute is suddenly very snappy. Users no longer have to wait for Overpass to respond and every click instantly produces a new love story.

---

You Can Use the Geocoder Too

Because the code is hosted on GitHub you can also drop my lightweight geocoder straight into your own map projects. All you need to do is include the script and point it to the cities.json file hosted on GitHub Pages. For example:

<script src="https://mapsmania.github.io/geocoder/geocoder.js"></script>

<script>

  const geocoder = new CityGeocoder("https://mapsmania.github.io/geocoder/cities.json");

  const nearest = geocoder.reverse(52.517, 13.388);

  console.log(nearest.name); // "Berlin"

</script>

That’s it - no API keys, no server calls, and no waiting on external services. Just a single JavaScript class, a static JSON file of cities, and you have a ready-to-go reverse geocoder that works instantly inside any JavaScript mapping library.

⚠️ One thing to note: the dataset is limited to just over 11,000 cities and towns worldwide. That makes it lightweight and fast, but it also means it’s not as detailed as commercial geocoders. It works perfectly if you only need to identify the nearest major town or city, but it won’t return smaller villages, streets, or individual house or business addresses.

🙂 The Emoji Map of Train Delays ☹️

In the classic 1990s management console game Theme Park, visitors’ satisfaction levels were shown through the use of small expressive icons - 🙂☹️. A similar visual cue has been used in Bloomberg's new mapped visualization of train delays on the New Jersey Transit commuter rail service.

In NJ Transit Is NYC’s Least Reliable Commuter Rail — By a Long Shot Bloomberg has mapped a series of smiley-face emojis onto a New Jersey Transit map. Happy, sweating, and sad emojis are used to represent trains that are on time, 10-30 minutes late, and more than 30 minutes late, respectively. The use of expressive smiley emojis on an animated transit map vividly illustrates the levels of delays on different lines during a particularly bad evening commute.

Bloomberg tracked more than 190,000 trains using real-time transit feed data from May through July 2025 to determine that NJ Transit passengers experienced major service disruptions at six times the rate of other commuters on its New York and Connecticut counterparts. The data revealed that there are frequent delays of 15 minutes or more, cancellations, and particularly long delays of 30 minutes or more on NJ Transit trains.

By translating raw delay data into an immediately understandable visual language, Bloomberg’s mapped emoji visualization makes the scale and severity of NJ Transit’s service disruptions as clear as a smiley or sad face emoji.

The Cold Case Murder Map

I'm not a huge fan of crime maps. I think in many cases they oversimplify or distort patterns of criminal activity. However, I find Japan's Unsolved Murder Cases project particularly compelling.

According to the map, there are at least 369 homicides in Japan (since 1995) that remain under investigation - because the murderer has yet to be identified. The map was created by The Asahi Shimbun, one of Japan’s oldest and most influential newspapers, in response to the 2010 Criminal Procedure Law. This law abolished the statute of limitations for crimes carrying the most severe punishment, such as murder. In practice, this means there is no longer a "time limit" on investigations for crimes committed after April 28, 1995.

To create the map the newspaper interviewed Japan's police departments about unsolved cases involving murder and robbery-murder that occurred between April 28, 1995 and December 31, 2023. The map itself presents a choropleth view of the number of reported cold cases in each prefectural police jurisdiction. 

Clicking on a highlighted prefecture reveals details of individual unsolved murders. Each case entry includes information such as the victim's age, gender, and the circumstances of the incident. Obviously these details can be explicitly gruesome (which may be why I feel a little uneasy about finding the map so compelling).

Via: weeklyOSM

Swimming With the Tide

English Channel Swim Tracking has released its 2025 map of English Channel swim attempts, charting the routes of all swimmers who have so far tried to cross from England to France this year.

The English Channel Swim Tracks 2025 visualization is very straightforward - just hundreds of line strings plotted on a single map. Unfortunately, the sheer volume of data makes it heavy to load - in fact, the map repeatedly crashes my laptop. However, despite this, the project is still extremely compelling, and I like it a lot. It has taught me something new, and in a way that only a mapped visualization could.

I had always assumed that swimmers crossing the English Channel took the shortest, straight-line route between the English and French coasts. One glance at the map was enough to disabuse me of that notion. As you can see, most swimmers set off in an easterly direction - almost as if heading toward the North Sea. Those who don’t tend to swim in a southwesterly direction, as though they were aiming for the Atlantic Ocean.

Swimmers rarely take a straight path across the English Channel because of the strong, shifting tides. The currents change direction roughly every six hours, so starting at the right moment is crucial. Many swimmers set off just before a tide shift, allowing the water to carry them in a generally westerly direction while minimizing sideways drift.

As the tide begins to change, swimmers adjust their course and gradually turn in a more south-easterly direction toward the French coast. This careful timing and navigation allow them to use the currents to their advantage, rather than fighting against them. The result is that the tracks often appear curved or counterintuitive on a map, with swimmers weaving in response to the tides rather than following the shortest straight line route between England and France.

The average time for an English Channel swim is typically between 10 and 16 hours (though conditions and swimmer speed can push this higher or lower). Because the tide shifts roughly every six hours, most swimmers experience at least two tidal changes during their crossing. This is why most tracks on the map show at least two major twists, as swimmers adjust their course to move with - rather than against - the tidal currents.