When mapbox first launched tilemill was when I first discovered them. I read about them on here, and went to check it out.
At the time I didn't really understand the value proposition. It was google maps without any of the useful API features like geocoding and routing. Plus they were using openstreetmap which at the time was lagging noticeably behind google data in my area [1].
Tilemill seemed nice, css to style maps. But google had a nice style wizard [2] that changed the maps on the clientside, no need to upload a custom tileset, which is what tilemill seemed to do.
To be honest I generally forgot about them. But then google started charging more for their maps, and I remembered them as an alternative. I still missed the flexibly google had for their map styles though.
However in the past week I've come to see them in a whole new light, in part due to this post, but also due to their vector map post [3] causing me to dig back in and play around. I now truly believe these guys are one of the coolest "underdog" startups I've encountered. They've been chugging along and have created some awesome advancements, or at least competitors in the online map space. I hope to continue seeing awesome work like this, those map images are gorgeous.
My only request is that they give a little more detail into how to create custom styled maps like they show in the vector blog post I linked to. The process of using their street maps with custom styles is a little hazy to me right now, although maybe I'm missing something obvious.
[1] It still does, but now only with minute details like service roads on the local university campus.
Any ideas of the detail level/resolution? Based on the "5 billion pixels" statement, it sounds like it would be something like 150-300m/px, which doesn't sound very high-res to me. For large scale maps it probably makes stunning images, but is it suitable for google earth -like zooming and panning around?
This release is down to zoom level 8, which is roughly 500 m to 1 km/px – what I think of as regional rather than local. You’re seeing mountain ranges and river valleys, not what you have planted in your garden.
Right now with this project we’re thinking about how to serve people making small- to medium-scale maps on topics like global environmental issues, shipping, travel, etc. At higher zoom levels (higher resolutions), our more conventional satellite and aerial imagery seems to be meeting people’s needs for now.
With the information they have and how they processed it seems like they should have a rough snapshot of each season. It would be remarkable to be able to click a button and change the landscape from winter-spring-summer-fall.
I'm starting to think we're stretching the definition of "beautiful" nowadays. I see it every other day describing js libraries etc and now this. The technical stuff behind this is interesting and sort-of impressive but "beautiful" is going a bit far.
Here we have a small showcase where a company took pictures of Earth from space - something that when described clumsily by second hand observers has brought tears to people's eyes - and their goal is to build a base layer that is the idealized, cloudless view of our planet in peak vibrancy.
A map of paradise.
And it's generated by a clever, possibly elegant, algorithm using tools anyone could use (aside from a massive server farm).
This is perhaps the most appropriately, unambiguously beautiful thing I've personally seen here.
It’s just barely clever enough that no one thought of it before. It has to be really simple, because the innermost loop (estimating how cloudy a given pixel is) runs several trillion times.
using tools anyone could use (aside from a massive server farm)
EC2’s m2.4xlarge (> 64 gigs of memory, 8 decent cores) is going for $0.14/hour right now. You could run ten of those all day and spend less than on dinner at a restaurant with cloth napkins.
I was the opposite. As soon as I saw that image of the UK, I really thought it was "beautiful", in the way that other people may think a piece of art is beautiful.
I find the maps they make really pretty, but the technical stuff doesn't quite deserve the hype. This seems like blue marble, which NASA did in 2005 (and every month since).
That said it should yield a dataset with twice the resolution of Blue Marble, which is cool. The size of a pixel will be 250m, i.e. a small town will be 2 pixel, as opposed to 1 pixel.
For high resolution imagery, you'd need to pay. There is 0.5m data out there, for about $5+ a kilometer, and I'd love NASA to dump a whole world data-set of this. But until then, I think Google and ESRI are safe.
Blue Marble was a big inspiration here, and we’ve had fun comparing notes with some of the people who made it. Blue Marble is, however, ten years old, and has a number of (justifiable) limitations. You can get a sense of these if you zoom in on its depictions of, say, the Niger Delta, or the northern tip of Greenland.
What has been lacking in the past is an open source type effort to make derived datasets available. There are numerous commercial projects that process public domain data and resell it, but do not publish methods or make the data available for free. The community tend to use NASA datasets with known limitations, or commercial data with high entry pricing. There is a distinct lack of cost-free domain specific derived data.
> I see it every other day describing js libraries etc and now this. The technical stuff behind this is interesting and sort-of impressive but "beautiful" is going a bit far.
Is that really what's being described? I don't think anyone is discussing the libraries this is written on when they say, "Beautiful." They're talking about the actual maps, which are things you could print out and hang on the wall if you wanted to.
I think it's awesome that they are using open data from NASA. I sometimes find myself skim through various open data sets and make list of possible projects/products with them.
That image of the UK reminds me that what I would really like is something that allows me to dynamically combine different maps and/or images with something automatically aligning the images/maps appropriately.
e.g. It would be great to blend that UK image, with a geological map with something like this image of the UK by night:
Look into Quantum GIS. It's an opensource tool that will let you do what you want (and more). A fair bit more flexible than Google Earth.
Most geographic charts and images are in formats that include a locations/scale/alignment. Mixing different coordinate systems (global vs uk local ones) needs the right settings. Good luck!
If you are looking to graphically align non-geostamped images, you should be able to figure out their corners by comparing them to a map with a known coordinate system, then use a .world file to specify that location stamp before importing them to Quantum GIS or whatever you use.
if the input data are georeferenced correctly then qgis will happily transform all layers to a common reference system if you check "on-the-fly" in the properties.
> they needed to download two thirds of a terabyte of compressed data. “We’ve got 30 to 40 servers pulling down data from NASA,”
40 servers to download 600 gigabytes of data? Something does not sound right here. If they wanted to avoid overloading the nasa pipes, they could have asked nasa to fedex that amount on couple of hard-drives or something. At this day and age bulk transferring a terabyte or two of data should not be a challenge.
Heh, yes, we seriously considered the Honda Civic full of external drives option.[0]
Some of our images were from NASA’s hot new GIBS[1] service, which was extremely fast. However, it’d only backfilled about 2/5 of the data we needed, so for the rest of it we were using a legacy endpoint that was launched around 2004ish and has some, let’s say, idiosyncratic caching and throttling systems.
Instead of setting up a special channel, we talked to them and figured out how to shotgun the downloads in a way that wouldn’t kill their cache or require them to give us special treatment. We thought of this on a Friday and wanted to have it ready when we came in on Monday, and it worked. Thus the somewhat blunt methods.
0. For any young persons in the audience, the reference is to “Never underestimate the bandwidth of a station wagon full of tapes hurtling down the highway.” from Tanenbaum’s “Computer Networks”.
My guess would be they wanted to uncompress and process it in parallel? I doubt it was a question of bandwidth—probably more a question of convenience.
I think it's fair to say that they thought it through. They also talked to NASA [1] so NASA would have told them if there was a better way (e.g. hard drive) than downloading over the net.
[1] “We called them up and said, ‘hey we’re going to hit you hard, what’s the best way we can do it for you?’”
At the time I didn't really understand the value proposition. It was google maps without any of the useful API features like geocoding and routing. Plus they were using openstreetmap which at the time was lagging noticeably behind google data in my area [1].
Tilemill seemed nice, css to style maps. But google had a nice style wizard [2] that changed the maps on the clientside, no need to upload a custom tileset, which is what tilemill seemed to do.
To be honest I generally forgot about them. But then google started charging more for their maps, and I remembered them as an alternative. I still missed the flexibly google had for their map styles though.
However in the past week I've come to see them in a whole new light, in part due to this post, but also due to their vector map post [3] causing me to dig back in and play around. I now truly believe these guys are one of the coolest "underdog" startups I've encountered. They've been chugging along and have created some awesome advancements, or at least competitors in the online map space. I hope to continue seeing awesome work like this, those map images are gorgeous.
My only request is that they give a little more detail into how to create custom styled maps like they show in the vector blog post I linked to. The process of using their street maps with custom styles is a little hazy to me right now, although maybe I'm missing something obvious.
[1] It still does, but now only with minute details like service roads on the local university campus.
[2] http://gmaps-samples-v3.googlecode.com/svn/trunk/styledmaps/...
[3] http://mapbox.com/blog/vector-tiles/