DiGEO: Navigating Digital Geography

I grew up loving maps. Our family’s World Atlas was my go-to book for hours of entertainment, more so than all the other books we had around. I made my first 3D map in 5th grade: a map of the Southeast Asian peninsula for a project on the Vietnam War. Notice the cafeteria tray!

My very first TInyMtn - 3D maps are fun to make.

As a child of the 80’s, I was also into computers. Around 2000, I discovered digital mapping and the USGS data FTP server and realized that mapping and computers were not unrelated. Since then, digital mapping has undergone an explosion in the availability, resolution, and precision of geographic data and tools.

This short talk is intended to be an introduction to geographic data, focusing on data sources and techniques that you can use right now, today, to make something beautiful for your office, home, or computer. But first…

Don’t be Afraid of Data

Apps, operating systems, tools, and global corporations all want to hide from you the fact that data exists. They think that all you want is “solutions.” If you’ve bought into that (and they’ve been pushing it hard since the dawn of the smartphone era), then you probably have something of a fear of data. So let’s remove some layers of obfuscation:

• Data is just ones and zeros (“bits”) which when grouped in sequences represent numbers (32 bits or 4 bytes) and letters (8 or 16 bits; 1 or 2 bytes), are stored according to some format specification (like JPEG), and saved on a physical memory device somewhere (RAM, SSD, disk drive, SD card).

• Contiguous chunks of data are called files, and are arranged logically in a hierarchy of directories called a filesystem. You have access to this, but it’s not always obvious how.

• Programs can open and save files and use the data therein, but only if that app was programmed to support that specific file format. For example, Photoshop opens .jpg but not .docx files. But you - the user - can look at the raw data for any file; it just won’t make much sense.

Types of [Geo]data

Since geographic (and hydro- and oceanographic) data represent properties and qualities of the Earth (or other planets), they must contain spatial (location-related) information. This “metadata” answers the question: what place on Earth does this data represent, and at what time? Since data is in files, the file’s format determines the amount, order, and specification of this metadata and of the data itself.

There are two main methods to represent data with locational properties: raster and vector.

Raster digital surface model and vector triangle mesh of the Evelyn Moakley Bridge in Boston

Raster

Think pixels - grids of squares, mapped to specific locations on the Earth’s surface The locational information is not stored for every data element, instead it is stored once for the entire grid (projection, corners, and pixel size), and the index of the pixels determines the place on the Earth represented by each data element.

Raster file formats that you may already be familiar with include common image formats like JPEG, PNG, and TIFF, and video formats like MP4, MOV, and AVI.

For geographic information, the GeoTIFF is very common - it stores the geographic metadata in headers just before the gridded data elements’ values, but it’s still just a TIFF image. You may also see “JP2”, which is a highly-compressed JPEG format.

The type of data stored in a raster file can vary - it’s not always aerial or satellite photography, but often elevations, land use, temperature, or other useful measurements.

Vector

Data for which each element also stores its position information is called vector data. PDF is probably the most common vector format file, followed by EPS (encapsulated PostScript) and a host of others. If you use Adobe products, you already know that Illustrator is the vector drawing tool and Photoshop manipulates raster data.

Vector data is very frequently used in Geographic Information Systems (GIS), where knowing precise property boundaries has significant monetary and legal weight. These “cadastral” files and other line-based vector data like roads, coastlines, and borders are often in a format called “Shapefile.”

Property and other boundaries are examples of line-type vector data, but for the artwork here, I use two other types of vector data: point clouds and triangle meshes. Point clouds are generally stored in LAS format and are the product of aerial or terrestrial laser-scanning - the data is simply three numbers for each laser “hit” and represent a point in space. Triangle meshes are the key data type used in 3D printing, are usually in STL format, and include not just 3D points in space, but connectivities among those points to form triangles.

Going Beyond Google Maps: Data Sources

It’s time to get active! Let’s download and view some geodata!

1. The USGS National Map

The USGS National Map download platform is the data source king -- supplying nearly all the data that I use for 3D prints of areas within the United States.

This repository’s most useful data sets for our purposes are: “Elevation Products->⅓ arc-second DEM” ,“Elevation Source Data->Lidar Point Cloud”, and “Imagery - NAIP Plus”. Draw a box, select a data source, search, and download - it’s simple and fast!

2. EarthExplorer

Though very similar to The National Map, USGS’s EarthExplorer portal allows discovery and download of current and historical global data, including declassified spy satellite imagery going back to 1960, single-frame aerial photos back to 1937, worldwide elevation data, line graphs, radar imaging, vegetation monitoring, and other datasets useful for global climate modeling. There’s a lot here, and it takes a while to peruse.

I most frequently use this site to get multiple high-resolution views of cities and landscapes from the “Aerial Imagery”->”NAIP” and “High Resolution Orthoimagery”.

3. The Copernicus Open Access Hub

The Copernicus Open Access Hub is my source for very recent high-resolution satellite photographs of Earth. The European Space Agency (ESA) provides this easy-to-use service with access to mind-boggling amounts of data.

You can download satellite images of your city with a resolution of 10 meters per pixel from as recently as today. It takes a little effort to extract useful full-color images from these approximately 1 GB data packets, but the effort is rewarded with wonderful color imagery of our beautiful planet.

Digging Deeper

To learn more about the tools that I regularly use to create 3D printable cities and mountains, read my other blog post: 11 Essential Resources for Your Digital Mapping Toolbox.

There is also a wealth of open-source software available for processing geo data, so if you want to dive deeper, consider trying QGIS or GRASS - both are open-source GUI applications for finding, viewing, manipulating, creating, and saving your own maps. If you’re no longer afraid of data or the command-line, the blog post mentioned above links to several of the tools that I rely on to create MiniCity and TinyMtn models.

I hope you’ve found something here that piques your interest -- even if you just want to enjoy exploring the wealth of fascinating and free data just a few clicks away.

Creating Art and Caring for the Planet

As humanity grapples with our environmental impact, each individual and organization has the opportunity to make smarter choices that minimize harm to the planet. 

To demonstrate how MiniCity does its part to promote a more sustainable, eco-friendly environment, we wanted to share the origins, materials, processes, and ultimate destinations of the components in our models.

In being one hundred percent transparent, we hope to gain your confidence in our commitment to not only deliver the most advanced 3D printed models, but to do so in the most responsible way possible. We are by no means experts in this field, so if you have other ideas as to how we can reduce our environmental impact or carbon footprint, we’d love to hear from you.

All of this is recyclable!

Components

MiniCity strives to use biodegradable materials, and whenever possible, purchase those materials from U.S. based suppliers. Here’s the complete list: 

• White PLA filament - filament and spools made in Indiana, USA; recyclable, biodegradable

• White wood panel - made in China, purchased in bulk; biodegradable

• Natural wood panel - birch wood, grown and assembled in Oregon, USA; biodegradable

• Glue - made in USA

• Gap filler - made in USA

Packaging

It’s easy to overlook the environmental impact of packaging supplies, and so we make sure to adhere to sound sustainability principles at this stage of the production cycle as well. 

• Bubble wrap (polyethylene) - made in New York, USA; easily reusable, recyclable.

• Polyethylene planks - made in Michigan, USA; recyclable.

• Product boxes - contain 40-45% recycled content; reusable, recyclable and biodegradable

• Shipping boxes - cardboard, some made in New York, USA with 80% post-consumer content; reusable, recyclable and biodegradable

• Tape - made in USA

• Rigid envelopes - chipboard contains post-consumer content; recyclable and biodegradable

• Paper - 100% post-consumer content; recyclable and biodegradable

Packaging a large MiniCity model of Kendall Square, Cambridge, Massachusetts

Shipping

Admittedly, we have far less control at this stage of the process. Still, we remain cognizant of our carbon footprint on this front.  

• We prefer to ship USPS, as they are the nationalized carrier in the USA, and have the largest regular delivery network (because special trips incur extra costs).

• When we ship with Etsy, we elect to buy carbon offsets for shipping.

• We are currently investigating carbon offsets for shipping for orders made at MiniCity.Art.

• Where possible, we order far in advance and use ground shipping, as long-haul trucking emits less pollution than air freight (though still more than marine or rail).

Final Thoughts

We are pleased with what we’ve done so far to reduce our impact on the environment, while still providing unique works of art for your home or office. But we know that there is more that we can do to reach our ideals. 

We hope you consider how your choices and actions matter to our interconnected world, and encourage you to prioritize practicality and beauty over consumption and waste.

Relevant links:

• Recyclopedia (from my home state of Massachusetts, but generally relevant): https://recyclesmartma.org/

• OECD report on pollution caused by various transportation methods: https://www.oecd.org/environment/envtrade/2386636.pdf

Your state or city likely has one of these - here’s ours!

Rendering 3D Cities

One of our most popular offerings is custom artwork. And since most clients inquiring about a unique and often complex art piece haven’t seen a MiniCity model in real life, it’s essential that we clearly communicate the process that will ultimately bring the client’s vision to life and accurately present what the finished product will look like. We achieve this via computer renderings.

The origin of the word “render” is Latin and old French, meaning “give back” or “translate.” In the context of computer graphics, to “render” means to “convert a virtual three-dimensional model into a two-dimensional representation” — i.e., to make an image of it, as if with a virtual camera.

An integral part of our production process is generating virtual three-dimensional models of cities. These “meshes” are composed of tiny triangles — the building blocks of 3D computer graphics. Once generated, we can “slice” them into layers for 3D printing or prepare them for rendering into a 2D image by applying spatial transformations and adding material properties.

Triangle mesh of Cloud Gate in Millennium Park, Chicago

In June, a potential client approached us with a request to create a 3D map of Chicago that would be larger than any other work we’d printed — a massive 5-foot-tall piece. Given the scope of the project, we wanted to reassure the client that the final product would match their vision and meet all their expectations.

To do this, we created a rendering using our virtual test room: a minimalist living room with an 8-foot ceiling, medium-grey painted walls, and two windows (the largest is in a wall to the left). Remarkably, when we showed up to install the piece, we noticed that the physical wall was nearly the same color as our virtual test room, and the light emanated from the same direction! We photographed the real model from the same angle as we did in the virtual rendering. Here are both images:

Radiance rendering of Chicago model

Photograph of actual model as installed

The differences between these two images (primarily attributable to an inevitably imperfect match of our test room with the actual room) notwithstanding, the final result is very close to what we, and more importantly the client, expected.

In order to make these renderings as realistic as possible, we leverage years of research in multiple scientific fields: lighting design, computer science, physics and optics, computational geometry, and computational physics. Radiance, an open-source software toolkit, combines this knowledge into a powerful tool for accurately simulating the transport of light through arbitrary scenes. It allows us to set up a virtual scene: to place walls, lights, furniture and our virtual artwork, and generate an accurate image of a potential installation.

When we talk about the “transport of light” we are referring to the phenomenon of light bouncing off anything in a room until it is finally absorbed. Darker objects absorb more light and allow very little to bounce, while brighter-appearing surfaces enable more of the incoming light to bounce. Light will bounce in one direction from shiny objects (called “specular” reflection), and it will bounce off of matte or flat-finished surfaces (called “Lambertian”) in all directions.

Imagine an interior space with light colored walls and ceiling surfaces; any incoming light (from an electric fixture, window, or skylight) will bounce around many times, filling the space with illumination. This is one of the main reasons why art galleries have white walls and architectural models are built with white paperboard.

Radiance can accurately compute the transport (bouncing) of light through arbitrary scenes. To illustrate this effect, below are several renderings of the same scene (the model above, focusing on the famed “Chicago Loop”) allowing light to make 0, 1, and 3 bounces. Not only is the scene getting brighter with each ambient bounce (as walls and floors become new “light sources”) but light penetrates deeper into the urban canyons created by the city’s numerous skyscrapers.

No ambient bounces

One bounce

Three bounces

Radiance is just one of the tools we use to create impeccably-detailed 3D-printed cities and landscapes. When you buy a 3D city from us, you’re buying artwork refined by decades of knowledge and experience, not only in physics and computation, but detail-centric design and model-building. We deliver the best, because that’s the only art we’d hang on our own walls.

Have a vision for a custom work? Let us know and we’ll work with you to make your vision a reality.

We Made a Huge 3D Terrain Map of Colorado

Here at MiniCity headquarters (my home office), we spend all of our time designing and building incredibly-detailed miniature models of cities. But years before we started making 3D printed cities, we were printing 3D mountains and terrain models at TinyMtn.

And so you can imagine my excitement when I got a call from BostonArt, a Boston-based art placement agency and gallery, asking us to build a massive 3D terrain map of the middle section of Colorado on behalf of their client--a bank in Colorado.

Working against a tight deadline, the project took shape inside the Artists For Humanity EpiCenter, a nonprofit in South Boston with the requisite space and equipment to house a project of this scope.

The finished map would measure at 9 ½ feet by 3 ½ feet--our largest piece to date. As I am writing this, the piece is on its way to the client, and should be installed next week.

Given the unique challenges arising from a project this size, I thought it would be worthwhile to share the step-by-step process from beginning to completion. It was an exhilarating learning experience full of challenges, surprises, and ultimately, satisfaction.

So, without further ado, let me tell you how we did it.

We start with the end: this is the piece in its final destination in Fort Collins, Colorado

Data Processing

The data type undergirding nearly all our projects is a digital elevation model (DEM). DEMs are raster (pixel-based) images where each pixel represents a square area of ground, and its greyscale value represents the elevation of that patch of land above sea level (dark=low, bright=high). 

DEMs of U.S. landscapes are available for download from the United States Geological Survey and for global landscapes at Viewfinder Panoramas or OpenTopography. For a detailed overview of these and other key digital mapping resources, check out our earlier blog post.

After downloading the data (for this project, we used Japan’s ALOS data from OpenTopography), we re-projected to the Lambert azimuthal equal area projection and cropped to the desired aspect ratio using GDAL.

The client wanted a slice of Colorado extending from the Eastern to the Western edge, but narrower in the North-South direction. We elected to include Boulder through Pueblo, omitting the San Juans, but including the I-70 corridor. Here is the DEM that we used.

This is the digital elevation model (DEM) of the middle swath of Colorado

Rendering

Prior to green-lighting the project, the client understandably wanted an idea of what such a piece might look like. So I turned to some of our Perl scripts that we use to generate 3D renderings of each of our MiniCity models with realistic treatment of light. Take a look at these intricately-detailed posters that we offer at our store. 

With some knowledge of the capabilities of CNC (computer-numerical control) machines, we decided that creating the piece from stacks of individual layers would be the most efficient. The alternative was to glue together a large and heavy slab of wood and carve the shape from that slab, like a sculpture. Several virtual tests showed that 16 layers allowed sufficient detail without excessive complexity.

Starting with the DEM as a 16-bit PNG image, I generated an outline of each of the 16 layers, converted each to 3D geometry in Gmsh, and ultimately rendered the collection of 3 million triangles on a virtual daylit wall using Radiance. As a side note, I’ve been using Radiance--a photometrically validated pseudo-radiosity raytracer--to generate realistic images for quite a while; it’s my go-to tool for complex visualization because it’s scriptable and accurate.

Radiance rendering of the proposed piece - this was the first indication of how awesome this would be

Sample

Since we had no back catalog of similarly large wood 3D models, we had to produce a smaller piece mirroring the final product in material and production methods--a model of a model, if you will. This was a great opportunity to test and refine our processes before committing to a major effort.

We chose a small portion of the full data, found 16 equal-spaced isolines of elevation, and leveraged AutoDesk Fusion 360 to convert those line drawings into toolpaths—commands for the Shopbot CNC machine to move the cutting bit. A big challenge here was to figure out how to get the Shopbot to cut our plywood pieces without slowing down on every spline curve (one easy-to-overlook setting in the driver software fixed this). In addition, we got to see what the polyurethane coat would look like on the Baltic birch.

The result was  a 10”x10” terrain map (image below) of the area surrounding Colorado Springs and Pikes Peak. It had a warm color and reasonable terrain detail, but there wasn’t much contrast between the layers. This persuaded us to consider a two-tone design, with alternating layers of plywood being stained to appear darker (as appearing in the rendering above).

A 10”x10” sample in Baltic birch plywood, finished with wipe-on gel coat

Design

Lessons learned from the sample (alternate layer coloration, number and detail of layers, etc.) guided the final design. We selected 16 layers, which at ⅛” per layer would extend 2” from the base board. In addition, the sheer weight of the 10” sample compelled us to develop a method to create rectangular hollows in the structure of the model: each layer would have material removed to reduce weight, but still allow sufficient overlap with the layers above to allow solid bonding during assembly. We used ImageMagick to perform the image processing and morphological operations to allow the generation of the layers’ outlines.

Below is a sequence of images for layer 4 from various stages in the processing pipeline. The bottom images were used to generate toolpaths for the Shopbot.

Multiple stages in the digital design of the layers, showing hollow regions for weight reduction

Cutting

We ordered scads of wood from Boulter Plywood in Medford, MA: 14 sheets of 60”x60”x1/8” Baltic birch plywood, plus a 4’x8’x1/2” sheet for the base, all ripped to 42” width, and they arrived the next day - yay! We chose Baltic birch plywood because of its strength, but mostly because it is made of many more layers than standard American plywood, showcased by the many exposed edges.

Using the above series of bitmap images, we generated (vector-based) svg drawings, then converted those to dxf in Inkscape, loaded into Fusion 360 software (which doesn’t have a Linux version—-this is the only non-Linux step in our digital pipeline), and ultimately generated toolpaths for the Shopbot CNC machine.

We first had to mill a pocket into a 4’x8’ MDF sheet to make our spoil board, which accepted our trimmed 42”x60” sheets of birch plywood (recall the final size would be 116”x42”). Most of our cutting was done with a 60 degree V-bit at 12,000 rpm at 50 inches/second feed rate. $200 Worth of double-sided masking tape from Scotch and McMaster-Carr held down the plywood, while the Shopbot turned wood into dust.

One of the early layers being cut from a 42”x60”x1/8” piece of Baltic birch plywood

As the pieces were pulled from the CNC bed, they were all labelled with layer number, a part letter, and even a part sub-piece number in the event they didn’t survive being pried up from the double-sided tape.

These are only a few of the myriad sizes of pieces we had to catalog, sand, and stain

All, and I mean *all* edges of *all* pieces were hand sanded (finger-sanded, to be honest) to remove burrs and smooth out the final appearance. Because there were so many thin pieces, some inevitably broke, but we ensured that they all ended up in the correct place.

Wear a mask, kids - dust is bad for your lungs!

Staining

Getting the stain right was a challenge. Undiluted stain goes on heavy and dark. But diluted stain differs in its effect depending on the application technique and time allowed to set.

We opted for a 1:3 dilution with mineral spirits (measured by the gram on a postal scale), application with a foam brush (to make it easier to get the edges), and 2-3 minutes set time before wiping. 

Nevertheless, all wood is different, and some panels came out a little darker or lighter than we expected. I’m a rocket scientist, so I wish everything was science, but, alas, wood is not. Fortunately, due to the 3D nature and alternating contrast in the final piece, any subtle variations are difficult to detect.

We organized each piece by layer number, intending to only stain the even layers. Despite this, we still stained some pieces from odd layers and forgot to stain pieces from even layers. Such is bookkeeping of puzzle pieces with power-law size distributions!

Testing two stain colors and three dilutions; sample Pikes Peak assemblies on the left

Staining area; a rag applies stain evenly, but a foam brush gets the edges

Assembly

Each layer had a guide sheet with the islands of land labelled and numbered. To place the pieces in their correct positions, we matched the labels that we wrote on the backs of each piece with the layer and number on the guide sheets. This took *much* longer than we estimated. Here is a photo of one of our guide sheets.

Where is piece 09-M?

We used at least two pounds of Titebond Original wood glue to attach all of the parts, smoothing the glue with a wood flange on large pieces, and with fingertips on complex and smaller pieces. 

Each piece was clamped or held down with sandbags (6 mil locking vinyl bags from McMaster-Carr and 100 pounds of playground sand from Home Depot) for at least half an hour before moving on to the next piece.

Gluing down the 4th layer - you can never have too many clamps!

Finishing

Boston mural artist Nate Swain lent me a hand with the finish - a coat of Varathane clear polyurethane, applied first with foam brushes, then touched up with a toothbrush. Yes, a toothbrush. This coating darkened the unstained wood and gave it a wonderful warm glow and also brought out a certain richness in the stained layers.

Nate did it with the foam brush in the shop.

Delivery

After a brief visit to the Artists For Humanity photo studio for some glamour shots, we wrapped the piece up with foam and stretch wrap to get it ready for shipment.

While ultimately the piece was crated, we only had to ensure its survival to the frame shop

BostonArt was gracious enough to send a truck to get the final piece, which, let me remind you, was nine and a half feet long, and weighed between 60 and 70 lbs!

I don’t normally nap during the day, but I did that day.

And there you have it: the step-by-step process of constructing a massive 3D map. The gratification of seeing the final product makes the process--entailing plenty of tedious manual labor, trial-and-error, and unforeseen obstacles--well worth it.

The finished piece on display in Fort Collins

Thanks

Obviously, I must thank a number of people who helped make this happen. Elizabeth at BostonArt was my go-between with the client, and a cheerleader when I needed one. Ryo at Artists For Humanity trained me on the Shopbot and gave me confidence when I didn’t have enough; Lorraine and Haidan at AFH facilitated the use of their work space, and John and Janna lent us their excellent photography skills. Nate, Suleiman, and Lauren showed up and sanded the many, many, small and large curved pieces that eventually went into the model. And, again, Lauren, my wife, encouraged me and supported me through the inevitable crises and breakthroughs that any big project entails.

If you’re interested in custom work by MiniCity.art, please contact us.

11 Essential Resources for Your Digital Mapping Toolbox

My fascination with digital mapping started over twenty five years ago when I first discovered the now-outdated SDTS-formatted third-arcsecond digital elevation models on the United States Geological Survey (USGS) FTP site.

Eager to combine the trove of data from the USGS with the new technology of 3D printing, in 2010 I launched TinyMtn, which today sells over 1,000 different models of the world’s most famous mountains, canyons, ski resorts, golf courses, and race tracks.

Since TinyMtn’s humble beginnings, digital mapping has undergone an explosion in the availability, resolution, and precision of geographic data.

Today, TinyMtn and now MiniCityArt are able to harness this veritable goldmine of data thanks to  a number of essential tools and platforms.

Here is a list of indispensable resources in my digital mapping toolbox.  

1. Google Maps

The most popular and universally accessible database of maps is still unrivaled when it comes to virtually scouting any location. 

Google typically incorporates new satellite photos and 3d scans shortly after they are released by governments and commercial providers, and augments these with its own data sources.

Most useful are the Terrain and Satellite views, which enable me to  scout new mountains for TinyMtn, as well as assess a landscape’s aesthetics. This is the first step in selecting a desirable build area to optimally present the geographic feature.

When generating MiniCityArt models, I use Google Maps 3D view to scout the city, identify unique building shapes, clarify various features, and gather key intelligence on the most interesting areas and popular neighborhoods.

2. Wikipedia

Cities and landscapes aren’t merely the sum totals of their physical features; they are the manifestation of thousands of years of dynamic history and culture. 

Overlooking a place’s unique history when designing a model can result in significant omissions and a rather humdrum final product that fails to embody an area’s essence. That is why the story of a place invariably informs my decisions about what to print. 

For example, if a historically or culturally significant monument, church, or building is located outside the downtown area on which the print is centered, I’ll consider reframing the print to include it.

As an added benefit, Wikipedia’s lists of tallest buildings in each city are thorough, and help with planning updates to each city.

3. The USGS National Map

The USGS National Map download platform is the data source king — supplying nearly all the data that I use for 3D prints of areas within the United States.

This repository’s most useful data sets for our purposes are: “Elevation Products->⅓ arc-second DEM” ,“Elevation Source Data->Lidar Point Cloud”, and “Imagery - NAIP Plus”. Draw a box, select a data source, search, and download — it’s simple and fast!

4. The Copernicus Open Access Hub

The Copernicus Open Access Hub is my source for high-resolution satellite photographs of Earth. 

The European Space Agency (ESA) provides this easy-to-use service with access to mind-boggling amounts of data.

You can download satellite images of your city with a resolution of 10 meters per pixel from as recently as two days ago. It takes a little effort to extract useful full-color images from these approximately 1 GB data packets, but the effort is rewarded with wonderful color imagery of our beautiful planet.

5. Geospatial Data Abstraction Library (GDAL)

Having access to comprehensive data is critical, but none of it would be usable without tools to convert and manipulate it.

GDAL is an excellent open-source library (suite of tools) that I use in many stages of my data processing pipeline, and I am extremely thankful for the programmers who have diligently developed these tools over the years.

Simply put, no open-source digital mapping toolkit is complete without GDAL.

6. Spatial Reference Catalog

It’s no surprise that presenting a spherical surface on a flat map or screen is intrinsically problematic.

A “projection” is a method for transforming positions on a globe (in degrees) into horizontal positions using measures of length (miles, kilometers). There are infinitely many possible projections, and map-makers must choose one when creating a flat map of our curved Earth. 

To aid in this effort, SpacialReference.org catalogs tens of thousands of these projections, coded using a variety of formats.

I’ll use either the Proj.4 code or the EPSG code for a desired projection, depending on the stage of my data processing pipeline. And I prefer UTM zones for their generality and meters for consistency.

I’m a metric person because I remember trying to perform thermodynamic calculations by hand in English units, and it upset me as much as English* food does. Measurements and food are both better when they are international — delicious and metric.

*Yes, the English use metric now, but they’re still largely responsible for confusing generations of engineering students.

7. ProjFinder

Sometimes — no, frequently — geographic data is distributed without embedded projection information.  ProjFinder tells us which projection the original geographer used to generate the data. Once I know this key detail, I can reproject and manipulate it alongside other sources of data. This terrific resource has resurrected a dead project more than once.

8. The National Grid

Did you know that the US has a “National Grid” — a general method for locating places within the United States?

Fed Maps is an ideal place to explore it.

Every place in the United States can be located using a code. Gateway Arch in St. Louis is 15S YC 450 788: 15 is the UTM “zone” (6 degrees of east-west longitude), “S” is the code for a band of 8 degrees of north-south latitude. Now the system switches to metric: “YC” is the longitude and latitude of a 100-kilometer sub-block of “15S”, and “450 788“ are hierarchical components of a 100x100 meter area within that. It’s an elegant system.

As you zoom in, you’ll see the unavoidable warping of what could be considered “North” and “South” lines — a side effect of the UTM projection system that vividly illustrates the difficulties of mapping a sphere. [Hey, listen flat-earthers: it’d be mathematically a lot easier if you were right. We don’t do all this extra work for fun, we do it because the Earth is a ball.]

9. Open Topography

This is an excellent source for both unique LiDAR (laser scan) data and an easy portal for downloading worldwide DEM data from Japan’s ALOS satellite.

10. Peakbagger

Attention all armchair mountain enthusiasts (and real ones)! 

A “peakbagger” is a mountain climber who methodically records their achievements, often with the aim of summiting all peaks in some category. Peakbagger — the site — is an invaluable asset containing lists and information about the world’s mountains. I typically use it to find the currently-published summit elevations for my new models, but it’s a great site to research new areas to print.

Peakbagger doesn't just tell you the location and summit altitude for each peak, it includes photographs, an interactive map, a hierarchical list of the mountain's ranges, and even a link to the peak's nearest higher neighbor (keep clicking until you get to Mt. Everest)!

11. ACME Mapper

If you want the name of a valley or basin, or a subpeak, or just want to explore beautiful topographic maps, head to the ACME Mapper site.

ACME Mapper lets you zoom in and out of a seamless display of USGS and world “topo” maps, giving you access to more details and a traditional map aesthetic that aren’t available on Google Maps.

—

While there are other tools and information sites that I leverage to print maps and landscapes, the 11 resources listed here are the most integral to the art and science of 3d printing. 

I hope you’ve found something here that piques your interest — even if you don’t ultimately want to make maps, and just want to enjoy exploring the wealth of fascinating data at your fingertips.