geobloggers

Geotagging becomes a real thing

Reverend Dan Catt — Tue, 06 Jan 2009 17:19:31 +0000

Apparently, as covered by all the other geoblogs, probably.

Oh and happy 2009, when my newest daughter reaches 6 months I’ll probably have had enough sleep to start blogging again, still in hibernation mode atm.

Flickr can haz (some) Shapedata

Reverend Dan Catt — Thu, 30 Oct 2008 22:13:59 +0000

Over at Flickr we’ve just turned on shapedata for various locations around the world. Aaron posts way way more over on Flickr’s Code Blog: The Shape Of The Alpha.

Now when you call flickr.places.getInfo if we have shapedata for that place, it’ll get sent back to you in the response. See Aaron’s post and the flickr.places.getInfo page itself for more information and example response data.

Since it went live I had a quick play asking for the shapedata for the States of America and it looks very much like this …

I did this by calling getInfo for each state in turn, asking for the response in JSON format. Using a canvas object over the top of the map I plotted each shape in turn. The results are pretty good.

Here’s France for fun …

… as I said in this post over here: Location, keeping it real on the streets, yo! …

“On a slightly more philosophical level, it’s a never ending process. We’ll never reach a point where we can say “Right that’s in, all borders between places have been decided”. But what we should end up with are boundaries as defined by Flickr users.”

These are not government borders, council borders or outlines from other agencies. These are generated based on the actions of Flickr users and how they choose to define their own locations, and which anyone can download and use themselves, with obviously the understanding that these are not government borders, council borders or outlines from other agencies, these are generated based on the actions of Flickr users and how they choose to define their own locations :)

Which is why it sometimes gets a little messy, see London Center Zoomed in …

… there’s various amounts of overlap, and sometimes gaps … in this case the big gaps are because I’m only plotting the top 100 “places’ in London, many of which fall outside of this view. The thin gaps between the shapes is due to the nature of the algorithm.

Anyway it’s kinda a fun stab at taking the local knowledge of the masses and turning it into something hopefully useful that anyone can use. Free shapedata for all … as long as enough people have taken a photo there :)

I’m pretty sure I’ll have to say on this topic again soon (oh and an example of how the canvas stuff all works).

Related: Here’s an image from Jim Bumgardner generated by grabbing geotagged photos tagged with the names of the lower 48 states: “Alabama”, “Arizona”, etc. and plotting them, allowing the shapes of the states to emerge …

(cc licensed http://creativecommons.org/licenses/by-nc-sa/2.0/deed.en_GB)

Where I’m actually living in augmented reality, Jefferson Airplane and what does this mean for photos.

Reverend Dan Catt — Fri, 24 Oct 2008 02:31:21 +0000

Here’s the breakdown, first I’ll mention augmented reality, then I’ll ramble a bit, then I’ll make a point, or not, depending on how the rambling has gone. First the Augmented Reality part.

I’ve downloaded RjDj onto my iPhone, it’s basically two things 1) Awesome, 2) Sound processing software that samples your environmental soundscape and plays it back to you mixed, looped, pitch shifted and generally fucked around with. It’s hard to explain so here’s a couple of photo/video things to try and explain.

First, this is me walking out of Montgomery Muni Station in San Francisco. It’s all very motion sickness inducing and next time I’ll try and make a steady-cam version. Wobblyness aside, this is now what I hear with the Eargasm scene selected as I walk around. The Eargasm scene is like being enveloped in your own Dystopian Future Eno Soundscape.

Here’s one from YouTube, a more tranquil outdoor soundscape …

To really have a go Download RjDj, insert in-ear noise blocking headphones, select Eargasm, walk outside, walk around. Imagine you’re in your own futuristic movie. Go out of your way to stand on metal grills, walk past people talking, a busker if you can manage it and near people talking, oh and near people talking. Eargasm only works well outside, Echolon is better in the office, imho.

I’ve been doing this walking to and from work for about 2 weeks now and have gotten used to hearing in this new way, it’s both exciting and calming. And I consider this augmented reality because the earphones I use blockout virtually all noise, if I’m wearing them with no music going on I hear just my own breathing. When RjDj is running the sounds I hear are totally passed through software and processed, I’m still aware of what’s going on around me, it’s just different, more enveloped in textures and so on. Eno, Eno, Eno.

You should really try it!

Now for the ramble part …

On my old laptop (and backup drives) I have all my music, purchased over the years, ripped to mp3s. iTunes tells me I have something like 60 days worth of music. Given that I can listen to around 6 hours worth a day, it means that if I play a track twice there must be another track that I can’t possibly have time to listen to this year.

So I bought an iPod touch and didn’t put all my music on it, I created smart playlists that’d pick 60 tracks I’d rated highly but hadn’t listened to for a while, mixed with other unrated tracks. A bit Genius before Genius came out. I thought that was neat, going from half a gahbillion tracks to just 60 at a time.

Then I bought a new laptop and figured I’d have to get round to copying all my music over. But I just haven’t done it yet, now with the power of last.fm I’ve, ummm, transcended the need to own albums and tracks and music and stuff.

I feel as though I’ve invested enough time in listening, rating and so on music that last.fm plays my music collection back to me pretty well. If I need something new I just listen to the radio station of one of my contacts, or my neighborhood and so on. Loving or Banning tracks as I go.

(Apparently I like female vocalists and electronic)

Curiously listening to neighborhood radio, or radio from my contacts, always ends up at Jefferson Airplane, no matter how many times I ban it. Maybe it’s a function of my age, the age of the people I know, and where I live. Could be worse I guess, could be Starship.

Last.fm generally covers me for music in 3G saturated San Francisco, but I still found myself in situations where I’ve had no coverage to pick up last.fm and no music on my iPhone :( Just Melvyn Bragg and In Our Time to keep me going. And at this point, I’m stubbornly not putting any music on my laptop or iPhone. I am now Post-Album.

And this is where RjDj fits in. It creates musical compositions on the fly and fortunately I happen to really enjoy ambient soundscape, other people I’m sure would totally hate it.

The only thing that would improve it for me, is if analyzed the current tempo, pitch, overall level and activity of the sounds, then based on my listening habits on last.fm, pull in snippets of music I tend to like, little echos of tracks I remember played here or there.

It’d be like your own personal version of the best album on earth ever; Chill Out by the KLF.

What does this mean for photos?

This is a bit of a jump, but here goes. I take a lot of photos, not as many as some people, and way more than others. Sets of photos are like Albums and one off shots like Singles. Just like last.fm they have views, tags, favs and so on. Just like my music collection I have more photos than I can possible look at, realistically anyway.

Now I’m not going to stop taking photos any time soon, but …

In the same way RjDj samples the sounds around me, and delivers them back remixed and processed, and heck at the top of this post I even recorded one so it can be played back later … old habits die hard it seems. What if the same happened with vision?

Here’s a 3D(ish) model of the City above Montgomery Muni Station …

My iPhone has GPS, it knows, sort-of, where I am, and which way I’m heading … with a little of a delay. But enough to give me a wireframe of what I’m looking at based on the 3D data it can grab, with at least the augmentation level that RjDj gives me with sound.

More mood board stuff …

Google Earth, with it’s 3D models and Street View Bubbles …

Rez (HD xbox) …

(images CC gamescoreblog)

flickr.photos.search API for finding photos within a radius of a point …

“Buttons” the Blind Camera … Read More “The camera memorizes only the time and starts to continuously search on the net for other photos that have been taken in the very same moment” …

Myvu Crystal Personal Media Viewers for iPod/iPhone etc (640×480 resolution, apparently not as awful as the older models) …

Does this mean that I’m nearly at the point where I’ll just be able to walk around, with my augmented sound and augmented vision? A HUD that superimposes a wireframe of where I am, resamples and processes what I see (I happen to like film effects, Black & White and soft vingnetting please) and mixes in my own memories of older photos I’ve taken at that point, while sampling and saving other views for the future and pulling in my contacts, friends and strangers photos?

In the same way that I no longer bother to carry around Albums of music, and listen to Tracks. Will I no longer need to actively record new photos, Have I invested enough time and source material for a system to recommend images to me and sample new ones for me. It’s not quite here yet, in-fact as I say this stuff, I can feel myself thinking “But I’ll always need to take photos, what about those special moments of the children growing up” and so on …

… but at the same time, there was a point when I though it was very important to keep all my music forever, for all time, to never loose a single mp3 in a drive failure. Yet now, I’m happy to not have even one on my mp3 player. I’ve reached the point where I’m no longer interested in cataloging, keeping and finding missing tracks, I have a soundscape that could feasibly play at me 24 hours a day. Even with access to my old laptop and music library, I just fire up last.fm and type in an artist or tag for the mood I’m in and let it go, banning or loving tracks if I really need to.

On the vision/photo front, this isn’t supposed to be useful by the way, in the same way that I use RjDj to make a walk more interesting or immersive. This is augmenting reality for passive/interactive immersive experiences to make walking to and from work (or for the joy of it) more entertainingly enveloping.

What I’d like, is to be able to walk along the beachfront from my childhood and be played snippets of sounds from when I was young, superimposed with old family photos, mixed in with music generated from the sounds of the waves, and my own and other peoples photos being pulled in from having been taken at the same point, but in a style I happen to like. All on top of what I can actually see, I mean, I still don’t want to be knocked down while crossing the road.

This is unlikely to happen for me now … going that far back, as I don’t have easy access to all those old sounds and pictures. But a lot of us are recording them for the future. When Modesty, Zachary and Isobel grow up, there’ll be more photos of them and events than they can possible get through. I suspect taking photos to keep as personal memories and sharing in the way we currently do will be terribly old fashioned.

But just because I no longer buy Albums, I still need musicians to make some music to feed the machine. And that’s the way I’ll be taking photos and videos and sounds and stories, and I hope other people will, so there’s enough to feed the realtime vision/sounds/mood processing machines of the future. It will, however, be very different, and probably fairly soon, and hopefully I’ll have something to do with it if I can :)

Europolis: What if Europe were condensed into one piece and combined as one cell?

Reverend Dan Catt — Fri, 29 Aug 2008 00:09:00 +0000

The lovely as ever We Make Money Not Art drew my eye to this piece by David Adjaye at Manifesta 7…

… the premise being “If Europe were condensed into one piece and combined as one cell, what would be left behind as residue? Two extremes: a very dense condition and a big void.” using a process of “… extract[ing] information from the capital cities of the European Union and condensed [them] into a single entity”.

I like the idea behind taking cities and trying to smoosh them together, joining them up in non-accidental ways. I’m sure there are other examples of fitting various maps together to form a whole, but they’re escaping me at the moment.

Regardless, photo attribution-share alike by we-make-money-not-art, [Via: The unusual suspects of Manifesta: Pirate Bay and David Adjaye]

Correcting Location Data … the Flickr way …

Reverend Dan Catt — Tue, 19 Aug 2008 01:47:37 +0000

It’s a little odd posting over there but now that Flickr does have a Dev Blog, it kinda makes sense to mostly post Flickr related geo-stuff there. And more comment here, even if I’m still stuck with an utterly dorky theme.

Anyway, recently I made a few posts about our new “corrections” stuff, you can read about it in more general terms on the main Flickr Blog: Introducing a new way to geotag and nerdy terms in the dev blog: Location, keeping it real on the streets, yo! with the first follow up here; Defining the boundaries we are all within.

In the Location, keeping it real on the streets, yo! post there’s a couple of paragraphs where I manage to nearly sum up why I think it’s important …

“For us, it’s a first small step into an experiment, and actually a pretty big experiment as we’re potentially accepting “corrections” from our millions and millions of users. We’re not quite sure how it’ll all turn out, but we’re armed with Maths, Algorithms and kitten photos.”

… and …

“On a slightly more philosophical level, it’s a never ending process. We’ll never reach a point where we can say “Right that’s in, all borders between places have been decided”. But what we should end up with are boundaries as defined by Flickr users.”

But what does it all really mean (to me)? Well, at a very basic level it means we’ve taken the descriptive geo-data on Flickr down to the next granular level. Instead of only going down to Town and City levels, we now go down to neighborhoods, and this is now reflected in the API.

Rereading Yahoo Woe (Where On Earth, that is) IDs but imagine it talks about going down to the next level will do the trick.

Before if you used flickr.photos.getInfo on a photo that had location information you’d get something like this …


	San Francisco
	San Francisco
	California
	United States

… but now you get (scroll right if you’re viewing this on the web) …


	Lower Haight
	San Francisco
	San Francisco
	California
	United States

The neighborhood node thingy being the new bit.

How’s this useful? Well in a hacky kind of way you can now pair up a couple of flickr api calls to do reverse geocoding …

Use the first one with a sacrificial photo you have write permission on (i.e. your own). Set it’s location with a Lat/Long and then go get the information for that photo to find out where Flickr thinks that photo was taken. There’s a couple of other ways of doing it, but that’s probably the easiest to understand.

The advantage to Flickr and the whole ecosystem that builds up around it of doing this is, as we feed back user submitted corrections into the backend, the neighborhoods Flickr assigns to a Lat/Long will slowly evolve over time to our user’s view of the world … and we have quite a lot of them, so in major cities this should be quite good.

This is hopefully converting a “database” view of an area (for want of a better term) into how real people on the ground view of an area.

On that subject, it’s also possible that users of FireEagle (now public), websites that integrate with FireEagle and users of websites that intergrate with FireEagle will also gain the benefit of potentially millions of people make updates to location, depending on how far back into WOE we can roll the data.

That to me is the magic part, people taking photos over here and saying “Oh no, that photo wasn’t taken in such-and-such, but actually here” can have a positive effect on someone’s experience with a totally unrelated website who happens to be using FireEagle to update their location. Suddenly that website gains the local knowledge from thousands of photographers :)

Well, that’s kinda the theory anyway, but it’s always interesting being able to take that kind of theory and attempt to put it into practice.

Map Shoes

Reverend Dan Catt — Fri, 01 Aug 2008 16:23:08 +0000

When I first saw that Zazzle were doing shoes my first thought was, obviously, ‘Quick, stick a map on them’ …

And since then I’ve been doing nothing but messing around with the designer and making map shoes of San Francisco (just north of Golden Gate Bridge)

Sadly they only do Womens and Kids shoes at the moment, and Modesty says she’d much rather have Kittens on her shoes than maps … so I guess it’s just me who finds them fantastic.

I was trying to think of something We Make Money Not Artsy about putting the streets your shoes pace along each day onto your shoes, or taking a little bit of home with you where-ever you go … but frankly I don’t have the prose for it.

My vote, Awesome!
My daughter’s vote: Meh.

Moving hosts … will be down and out for a couple of days.

Reverend Dan Catt — Tue, 22 Jul 2008 18:37:49 +0000

So I’m finally tired of Yahoo! Small Business’ WordPress blog hosts. It just doesn’t seem to work. The database falls over, connections don’t work, the log files fill up, or whatever. The WordPress versions which are supposed to auto-update don’t seem to leaving security holes, and so on.

Therefore I’m taking the easy route out and moving over to being hosted on WordPress.com. Which means I can hopefully spend more time actually writing posts than battling with my hosts install.

Over the next few days I’ll be porting my old posts over (hopefully)

iPhone not writing Location EXIF correctly?

Reverend Dan Catt — Thu, 10 Jul 2008 21:34:41 +0000

Ok, so everyone’s playing with the new iPhone 2.0 software, and taking photos, and embedding the location information into the EXIF.

Well early adopters in San Francisco are anyway, and here’s an interesting thing about San Francisco is that it’s at the same latitude as a bit of the Yellow Sea, just off the coast of China.

Example: San Francisco is 37.7 latitude, -122.4 longitude which puts it firmly in the Western Hemisphere. The same latitude but 122.4 longitude you end up just off the coast of China.

The Spec for GPS location in EXIF says that you’re not supposed to have any sign information on the latitude or longitude, so it’ll always be positive (i.e. the North Eastern Hemispheres) in the "GPS Latitude" and "GPS Longitude" fields (although some apps do add that information) The North/South East/West bit is supposed to be in the "GPS Latitude Ref" and "GPS Longitude Ref" fields.

But it seems at first glance, that the iPhone 2.0 software misses these fields out when writing EXIF data to the photo, so while everything else using Core-Location on the iPhone is super happy. Photos appear to be missing which Hemisphere they’re taken in.

And those photos taken in San Francisco, well, they think they’re in China … for the moment.

EXIF from iPhone

GPS Latitude    37 deg 45' 36.00"
GPS Longitude   122 deg 25' 48.00"
GPS Position    37 deg 45' 36.00", 122 deg 25' 48.00"

Missing EXIF Fields

GPS Latitude Ref        North
GPS Longitude Ref       East

Can anyone else confirm that there’s no North/South, East/West information in the EXIF, or is it hidden somewhere else and I’m just missing it?

Here be Updates:

Photos from an undiscovered country
iPhone 2.0 Geotagging (Also putting someone in China)

Seems like there may be differences between first edition iPhones running version 2.0 software and the new 3G iPhones running v2 software. And also a difference between photos geotagged when the GPS is running outdoors to when they’re geotagged from WiFi location only.

iPhone has GPS (confirmed)

Reverend Dan Catt — Mon, 09 Jun 2008 18:41:02 +0000

I mean we all knew it anyway, so I guess I’ll have to buy one now.

Now lets see how it actually writes the information to the EXIF fields ;-)

Google Map Photos, Redux.

Reverend Dan Catt — Wed, 04 Jun 2008 06:55:44 +0000

A few weeks ago I commented about Googles new photo overlays, in which they pre-bake Panoramio photos into tiles. I said I was unsure of why they didn’t have a more even distribution of photos at the global level.

Well it seems that (totally co-incidently I’m sure) a more even spread has now happened. Although I have no idea when this change came into effect, I see no mention on any of the Google/Panaramio blogs. Anyway, this is what the global level used to look like …

Old:

… compared to how it looks now …

New:

The new is definitely a marked improvement. There’s a lovely even distribution of photos, and even a little artistic flair in having a select few larger photos. It certainly feels like the larger photos are picked out as being more photogenic too, although I have no idea if this is actual true, and if it is, if it’s editorial scoring or some other method to devise photogenicness :)

The spread happens across all levels. Again a comparison between the North edge of San Francisco from a few weeks ago to now…

Old:

New:

Again, the difference is obvious. And while it’s definitely better, I guess I maybe hard to please, but it’s still not quite right (imho) when you’re zoomed in at this level. A mix between the two in this case would be great.

While the new has the photos less obscured by each other, the older version does a better job of showing where the overall distribution of photos is. You can see the clusters around the Northeast edge and over the two bridges. The algorithmic spread introduced into the newer tiles has artificially effected the distribution of the number of photos taken in different areas.

As an example it looks as though there’s just as many photos taken in the water just off the North edge of San Francisco as on the land just under the North edge. While on the older tiles you can see this isn’t quite true.

There’s probably a happy medium somewhere between the two. Like the strength of the clustering effect being proportional to the zoom level. The more zoomed in, the more photos cluster, highlighting dense photographic areas. The more zoomed out, then less clustered and more evenly spread, giving a good visual sampler of large areas of the world.

Anyway, good to see that improvements take place, and I’m still interested to see what “themed” pre-baked tiles would look like, such as “night” or “sunset” or “celebration” or what have you.

Look Around Photos in 3D (ish)

Reverend Dan Catt — Tue, 03 Jun 2008 23:53:12 +0000

By some strange quirk today I came across this great way to, well, look around photos taken at a single location …

… by using Creative Commons photos from Flickr. Although I’m sure any repository of geotagged photos would do.

Far more information in the creators blog post over here http://openphotovr.jottit.com/re_panoramio. This screen shot taken from http://openphotovr.org/#a362

(Sydney opera house)

[via: TechCrunch]

Geomancering, the word that doesn’t work :(

Reverend Dan Catt — Fri, 30 May 2008 15:29:19 +0000

I love the word Neuromancer, from the Gibson book. For me it works on a several levels and conjures up all sorts of evocative images, it’s a near perfect word. It has that cross between someone who studies and controls the dark arts of Neurology, and a “romancer” of Neurons. That lights up the Mondo2000 part of my brain.

So approximately once or twice every 6 months I think, “no wait, cool, what about Geomancer?” Someone who dabbles in the art of NeoGeography?

And then I remember that the word is already taken.

Which sucks.

No doubt I’ll go through all this again in around 6 months time.

(as an aside, are my comments working? They seem fine to me, but I’ve had a couple of people say they get a blank page)

Google Earth in a browser (sort of), Scriptable, a quick peek and poke.

Reverend Dan Catt — Wed, 28 May 2008 17:38:33 +0000

Radar points out that you can now get a ‘Lite’ (my word not theirs) version of Google Earth as a plug-in for IE and FF on Windows only (boo!), more information over here. Which is, you know great. More-so this is the really exciting part, for me anyway …

Google Earth, as an application is hard to script and make-do-stuff^(tm), a very cursory glance at the Google Earth Plugin shows a different story.

First, make sure you have Firefox with Firebug installed, or just follow along with me, with pictures.

Then go here (on Windows, blah blah) to see an example page. You’ll probably have to install the plugin and restart the browser. A quick view of the source-code gives us a few hints as to what to do.

Open up the console in Firebug and type “ge” (without the quotes) to see what javascript thinks the “ge” object is, you’ll probably get a line afterwards like this …

[xpconnect wrapped (nslSupports, IKmlLookAt, IKmlAbstractView ...]

… clicking that line will expand the whole object into the Dom, which will look something like this (only more) …

… See, see all the lovely methods! I generally find this (initially) a far more interesting way to discover what methods can be used etc. when playing with new stuff, especially as the API docs are normally not fully there right at the start. The one I’m particularly interest in is the “ge.createLookAt” method. Let’s use that, again in the console type these 3 lines …

var la = ge.createLookAt('');
la.set(51.503226, -0.118876, 100, ge.ALTITUDE_RELATIVE_TO_GROUND, 0, 0, 4000);
ge.getView().setAbstractView(la);

They will, in turn…

Create a LookAt object
Set the target of the look at object, in this case the London Eye, in London
Tell Google Earth Plugin to set the view to the LookAt object

And that’s it. Whoo!

Or you could, you know, use the specially built sandbox to play with the methods that Google have, but where’s the fun in that?

Yes I know that’s not much, but it’s a damn site more scripting than I’ve been able to easily do with Google Earth the App. Of course this is only just touching the tip of the iceberg, and I’m sure we’ll be seeing a number of more interesting examples come out over the next few days. I suspect the first few will just be laying KML files over the top, nice but dull =p. But hopefully after that we’ll start to see more code where Javascript is controlling how the globe moves around to give the user a more semi-interactive experience.

Tie in a little Ajax and there’s no good reason why at the very simplest you can’t have one person controlling the view for several people sitting at different machines (and yes I know you can sort of do this with Network Links in Google Earth, or normal Maps, but this has a certain immersive element to it).

Taking us all slightly closer to Snowcrash :)

Oh, and here’s a dump of the LookAt Object …

… and this is the stuff I was typing into the console to test things out, where I got the location of the London Eye wrong at least twice!

Once more, I know this is all very trivial, but it makes me smile. Bring on the Google Earth Javascript games.

Yahoo Woe (Where On Earth, that is) IDs.

Reverend Dan Catt — Mon, 12 May 2008 20:52:45 +0000

Roll up, roll up, roll up, get your WoE IDs here …

http://developer.yahoo.com/geo/

… a jolly nice step in the right direction.

Yahoo have opened up their geo database (which is pretty good btw) which is far more awesome than I’m going to make it sound in this blog post. I already did a little bit back here. But a quick call out to these two functions that’ll try and find you WoE IDs based on string input…

… the first API call gives, 36240 as the value for Stoke on Trent in the UK, my old hometown. With that WoeID I can plug it back into these other API calls to get various useful information back …

We use WoE IDs over at Flickr, again you can read a little more about that near the end my terribly long previous blog posts about twitter, APIs and such like. A quick recap is that we have these WoE related APIs …

The ones that probably compliments the APIs over at developer.yahoo.com/geo are flickr.places.findByLatLon, which will turn a Lat/Long into what Flickr believes is there, more on this in a second.

And flickr.places.resolvePlaceId, if you plug the value of 36240 into the API explorer there you’ll get this xml back …


	
		Stoke on Trent
		Staffordshire
		England
		United Kingdom

Where you can see the same woeid, the URL to get to the Places page for Stoke on Trent, and the parent hierarchy flickr uses.

You can also plug the woeid into the flickr.photos.search to get photos back for just that area. A quick example of why that is useful is California, there’s no way you’d want to find photos for California using the bounding box, as it covers a couple of other states …

… but instead you can use the woeID for California (2347563) which deals with all the bends and kinks of CA. You can also throw that ID into the Places URL, like this … http://www.flickr.com/places/2347563, although you’ll probably want to pass that through flickr.places.resolvePlaceId first if you want a pretty URL.

Differences between Yahoo geo API stuff and Flickr

Over at Flickr, we only use specific ‘levels’ of geo information, such as city, region, state, country, while the APIs over at Yahoo will spit out far more levels in-between the ones Flickr uses, as well as deeper down to neighborhood levels, which Flickr doesn’t do (yet).

Just because you get a WoeID back from Yahoo, doesn’t mean we’ll have photos for that specific area, we’ll probably bounce you up to the next largest places we deal with. So our parent hierarchy will have less steps in it that the ones you get back from Yahoo’s geo stuff. It also means our find by lat/long will only go down to town/city level and not precise neighborhood level.

We also don’t do photos or Places pages for Landmark WoeIDs, taking their example of Sydney Opera House which gives you an ID of 28717584 and throwing that at Places will give you no photos (maybe one day), although you can bounce up to Sydney.

Anyway, it goes beyond photos at Flickr, it’s just a really useful way that you, a developer can key off an ID for a place, that someone else, somewhere else can also key off.

If you use the ID 2347563 for California, and they use the ID 2347563 for California. And one, or both of you, publish your information with that ID, then you, they, or other people know that you’re both talking about the same place and match that data up.

Which is nice.

[Update 1: If you're coming here from the CNET news article my talk isn't actually about the location platform, its about Flickr photos and location :)]

[Update 2: For more official stuff about the Location Platform, you should probably see the Yahoo Local & Maps Blog post Abstracting Spatial Relationships with the Yahoo! Internet Location Platform where they use phrases such as "unambiguously", "permanent" and "language-neutral", and sum everything up far better than I]

[Update 3: and if you haven't yet checked it out, I still think its worth reading my blog post "Twitter API updates, FireEagle and new Flickr API fun" as an example of what to do with woe IDs]

Google Map Photos and SuperGeotagged

Reverend Dan Catt — Fri, 09 May 2008 18:20:20 +0000

There’s been a bit of coverage about Google’s new addition to maps.google.com, with VentureBeat probably adding the most opinion about it. Google’s video has it pretty much covered …

… Ok, so just a touch deeper at what I think is going on here.

Google snapped up the rather wonderful Panoramio a while back, a photo site built on the raison d’atra of location. And not knowing the full details I believe that Google/Panarmio have a cycle of someone or something, or a combination of both of those, looking at all the new photos (since the last cycle), and deciding which ones will make it into the next Photo layer on Google Earth. Or rather I suspect they decide which ones won’t, but the principle is the same.

The new crop of photos, and removal of old ones that have either gone from the site, or someone has asked to be removed, are then converted and stored in a Layer for Google Earth. So when a version of Google Earth is released, it comes with the latest “Data Pack” of Panoramio photos. There’s no actual ‘live’ searching of the Panoramio site that I’m aware of.

Now, at the same time as making the layer, Google can also prepare the data for Google Maps, let’s take a quick peek at San Francisco …

… here’s a similar view from SuperGeotagged (using flickr photos, more on that later. I also think they could slightly smaller images at this zoom level) …

This is done by loading in an additional set of transparent tiles over the top of the map, this is tangently covered in my previous blog post Paul Smithâ€™s thoughts about maps, it’s worth scrolling down to read his comment at the bottom.

Here’s an example of a transparent tile from both sites …

… Google’s tiles are served from their own servers, and SuperGeotagged supply their own slightly less ridged looking tiles.

SuperGeotagged: You can read more about the SuperGeotagged over here: SuperGeotagged - Every Geotagged Photo. Which as it happens isn’t every geotagged photo, just everyone with the “geotagged” tag, around 1.5 million of them. I have no idea how long it took Sean to render the tiles, but he needed to grab the thumbnail for each photo, so I suspect a long time.

Google on the other hand could possible render the tiles on the fly, if you compare the two images below, the first from Google and the second from SuperGeotagged, it looks like Googles coverage is quite sparse, which is isn’t. Actually they seem to have just deciding to render only a certain number (sorted somehow) of photos at a time, for if they are using the same photos in maps.google.com as they are in Google Earth, they should have over 3x as many photos represented here than you see on the SuperGeotagged map.

The difference you see below is fairly dramatic …

… so either Googles tiles are being rendered on the fly (and cached) and only the ‘top’ 200 or so photos are being used to keep thing quick. Or they are all pre-baked, in which case I don’t understand why they don’t go for a more even distribution at the global level. There’s seems to be no reason for google to render the tiles on the fly as no dynamic search is actually being done.

Which brings us to actually how useful is it to display as many photos as possible on a map?

Well with pre-baked tiles, it’s not actually possible to perform searches for images and view them in this style. I can’t just say to either Google or SuperGeotagged, show me photos tagged ‘beach’. SuperGeotagged would have to ask flickr for the photos via API searches and then render the tiles. Google have it slightly easier in that they can have the data on hand, so throwing enough machines at it makes it possible, maybe, sorta.

Each of these sites could pick a theme if they wanted too, “Windmills around the world” and then Google could show them on Global Windmill Day, or something, but then that’s editorial controls, which for the record is just fine with me.

So what are we left with? Well the photos are good for showing us the distribution of where people take photographs (generally where people sell cameras as it turns out), and I believe SuperGeotagged does a better and more artistic job of this. It’s really showing you, “yeah, there’s piles of photos here”.

Google on the other hand, leaves me slightly cold. When you first “Explore the area”, you basically get 8 photos and 2 videos placed live onto the map (the same way that Panoramio do -video), which is great, but at the same time I can’t actually search for stuff, like “coffee”. Or rather as soon as I do, I lose the photos. It being Google, I feel that I aught to be able to search for photos (and the resulting photos to be from the internet as a whole, not just Panoramio, but that’s a whole different kettle of fish).

Then when I click to view “More photos” I get thrown into pre-baked tiles mode, which is great for getting a feel for a place (which you can do with straight image search), which looks like it’s trying to be useful (in a way that SuperGeotagged doesn’t pretend to be), but isn’t so much. I can’t actually click on all those photos of the Golden Gate Bridge because they are all on-top of each other.

Anyway, this is starting to sound like a rant, so I’ll end with this.

I know the people at Google are smart, so I expect something a little smarter. Just because you can throw thousands of photos on a map, doesn’t mean it’s a good idea, no matter how pretty it looks. Unless you’re offering it up as a pretty-thing (like SuperGeotagged), not a search-thing (like what Google is known for).

(and I’m not saying this because SuperGeotagged are using Flickr Photos, in theory with Panoramio’s API they could use ‘Googles’ images, and I’d still prefer the way SuperGeotagged have done it than Google, because of how it’s presented).

Oh and yes I know, its easy to lay out criticism like this, we (flickr) are far from perfect ourselves, with our own collection of problems, but I’m actually in a position to do something about these here. With Google I can just poke, prod and nag, they can always poke, prod and nag back at me next week at Where 2.0 :)

Geotagging Goes Mainstream …

Reverend Dan Catt — Fri, 02 May 2008 00:35:11 +0000

… you know officially!

Before today we were like totally niche, but now that Microsoft has declared “Geotagging Goes Mainstream” I guess we’re out in the spotlight now. Next thing you know we’ll have major newspapers writing articles about geotagging and everything!

All kidding aside this looks like a pretty good tool. I say looks as I’ve not had a chance to play with it yet, being at home with just a Mac to hand, I’ll give it a whirl tomorrow and see how it goes. One of the nice touches that we’ve all talked about, but I’m not sure if anyone really got round to doing well is having a time-offset-slider, I think you can see it in the screenshot below …

screenshot courtesy of microsoft

… where you upload your tracklog and it matches the time with the timestamps on your photos. The theory being that if you’ve forgotten to adjust your camera for daylight saving, you can shift all your photos along the track until they’re right.

They also suggest that when you drop a photo on the map, be it by hand or automatically, it’ll also add location tags to the photo for Street Address, City, State, and Country. I’ve no doubt that they can get the city right most of the time, but I’m interested how they reverse geocode the lat/long to a street address on a global level.

Which reminds be I should get round to writing that “why reverse geocoding is hard” post at some point, maybe Microsoft have solved it, who knows? Like I said, I’ll take a look tomorrow with some of our known problem areas :)

Anyway, from my point of view, this is a great thing. If a user uses it to catalog their photos with tags and so on in the “metadata” and does the whole geotagging thing too and then uploads those to flickr (or to Microsoft’s photo sharing site thing, do they have one of those yet?), so that it extracts the tags and geolocation stuff, then that makes me happy. Anyway the more people that find geotagging easy, the better it is for everyone else.

And the more photos they can pull back out of Flickr for Photosynth ;-)