Home | Blog | Brainjam | Publications | People

World Explorer: A JCDL best paper!

Last night, we discovered our work on TagMaps / World Explorer won best paper at ACM/IEEE-CS Joint Conference on Digital Libraries (JCDL). Rahul’s in British Colombia where he presented the paper, World Explorer: Visualizing Aggregate Data from Unstructured Text in Geo-Referenced Collections.

Read the paper, check out the demo and many shouts to Shane, Mor, Rahul, & Jeannie. Fantastic!

Oh and thanks Rahul for letting us know via this Flickr photo.

Hack with Y!RB in London

What’s in a location? Our Tag Maps, ZoneTag, and Zurfer applications all involve some location-based services. In Tag Maps, we have a bounding-box query that returns a list of automatically-computed landmarks/attractions nearby. In ZoneTag, a location query returns a list of tags that might be relevant to the user at that location — perhaps useful to tag a photo they have taken there. ZoneTag also helps with Cell Tower tracking: give the ZoneTag service a cell tower ID, and it might actually tell you where in the world that cell tower is. Zurfer can also offer a few useful services - but these are not yet released.

These services are already available on the Yahoo! Developer website. If you happen to be located in London, though, and you are a hacker, and you are planning to come to Hack Day… then you might be hearing about all of these Y!RB web services first-hand this weekend.

Plus, FireEagle!

CrowdScapes on the Go from ITP

Mr. Bukhin likes to make us think about places and images. From his adventures into memory and implicit capture at Waymarkr (with Mr. DelGaudio), Mike has been raising more question than answers in his ITP projects. That’s a good thing.

Recently, Mike created CrowdScapes, his final ITP thesis project. CrowdScapes proudly utilizes our own TagMaps data API (as well as the Flickr APIs, of course). Mike describes the project this way:

My mobile application, CrowdScapes, lets a user explore a neighborhood through the crowd’s eyes. CrowdScapes value is that it leverages the critical mass of a large community of photo takers and sharers, not just a small subset of power users using a custom application. CrowdScapes can give places a user passes through everyday but doesn’t really consider a new life and a new possibility. By letting users step outside of themselves and consider what a location means to others, CrowdScapes can give new insight into a place.

The view CrowdScapes provides moves around two core pivots. The ‘familiar view’ shows localized photographs by a participant’s most used tags, it shows the current location through the lens of the participant’s interests. The ‘strange view’ shows the most popular tags and their respective photographs as viewed by the global Flickr crowd.

The TagMaps data is what helps Mike generate the `strange view’.

You can try CrowdScapes on your phone (via WAP) at http://m.crowdscapes.com. In a couple of days, you will see something new from us that will touch on very similar concepts. Start getting excited.

Oh, and congratulation, Mike, for passing your thesis defense!

TagMaps / World Explorer - A History Lesson

(News item first: Today we released an update to the TagMaps embed object - developers, check it out here.)

In late 2002, I first started working with geo-referenced photos as part of my PhD. We tried to imagine a future in which cameras have built-in location capabilities. What could we do to make the life of everyday photographers better? One of the first ideas that came to mind was based on the fact that location information allows you to correlate metadata (such as descriptions or tags) between media objects from different users. Based on this intuition, we developed a system called LOCALE (paper here) that allowed tourists (taking the Stanford campus tour, in our case) to snap geo-referenced photos, upload them, and add captions to the photos. The system them extracted terms from captions attached to these geo-located photos. After we had enough data (from over 50 participants) we could see that certain terms “float to the top”, or are more popular, in different campus locations : Hoover tower, Memorial Church, Fountain, etc.

While we ended our experiment with roughly 1000 photos, all taken on Stanford campus, we dreamed that a much larger dataset, covering the entire world, will be available one day. Such dataset can take us from LOCALE to “GLOBALE”: a system that provides a complete “picture” of the world, understanding and visualizing what every location is about. But where would we get the data? What would be the algorithm? How would one present this information to the user?

Four years after LOCALE, most of the answers for our imagined GLOBALE are finally here. It took the ingenious incentives system of Flickr to get people to contribute the data - geotagged photos and associated text tags (currenly over 5000 photos are available from Stanford campus). It took the researchers and interns of Y!RB to analyze the data and devise the algorithms in a way that allows extraction of meaningful patterns, and serving the data in real time without delay. It took some more clever ideas from the same research team to come up with the visualization and interaction techniques. Finally, it took some help and tips from our designer friends in making it all look good.

Let’s take one paragraph and acknowledge the people who made all this happen. The core Y!RB Media in Context team: Shane, Rahul, Simon, and Jeannie. The intern: Alex. Ideation: Marc Davis. Help with algorithms/methods: Tamir Tassa. The Flash SWAT team: Jeffery and Joe. Oversight and and adult supervision: Ellen. And, of course, the Flickr folks who created this wonderful world of photos, and the Flickr community who contributed, and continues to contribute, such compelling and rich data.

Oh, and by the way, if you still want to use the LOCALE data for research, give me a shout.

[update: I just learned that Arun Qamra at UCSB have been using the LOCALE dataset for his ICME 2005 paper. Cool!]

Introducing: World Explorer and TagMaps

Can we automatically extract information from Flickr geotagged images to create a rich visualization of the world we live in? The answer is: you bet.

Introducing: World Explorer. While the amazing PhotoSynth (from Microsoft!) allows you to explore a specific landmark in depth, World Explorer opens a window to explore the entire world through the eyes of the users of Flickr.

We use the public geotagged photos contributed to the world by the lively and active Flickr community, and the tags associated with these images, to help you explore the world like never before. Our system automatically extracts the tags that are relevant and representative for each map region or zoom level — and connects these tags to the photos that represent that area.

Ok, enough with this high-level blah blah. Start exploring! Find out what’s important in Paris. Check for what you could see if you go north of Golden Gate Bridge in San Francisco. Discover whether there is anything interesting on the west coast of England. Decide whether you should go for vacation in Maui or Kauai. Make serendipity lead your way around your own home town (Yoda in San Francisco?). Or simply explore Africa. If you discover something interesting, don’t forget to let us know right here.

To add to the fun, we also introduce Night Explorer and Trip Explorer. Night Explorer visualizes Flickr night photos in the same way, giving you a view of how Flickr users see the world after dark. Trip Explorer, on the other hand, is not based on Flickr at all - but rather, on the contribution of Yahoo! Travel’s Trip Planner users. What are the interesting Trip Planner items in Vermont? What do people visit (ahem, plan to visit) when they go to Austin, Texas? What are the non-obvious locations to visit in Los Angeles? It may be interesting to compare the three different views (Night, Trip and World Explorer) for the same map area.

Still with me? Not exploring yet? Well, good news: you can embed a version of TagMaps / World Explorer on your blog, web page, or anywhere else. Just like we did here:

As you can see, this embedded element is set to take you to Yosemite Valley in Yosemite National Park. When you put it on your blog, you can choose whichever place you want to start with.

Still reading? You should be really be exploring by now. If you are reading this far, you must be a developer. Well, we have something for you, too.

World Explorer is using two main components. The first is TagMaps, a Flash/SWF object that visualizes tags (i.e., text terms) on a map. The second component is a data APIs - an API that provides the tags to display, including the location and size for each tag. Both of these elements are available for developers to use. Mash it up! you can use both elements together to have a version of our World Explorer on your own page. Or, you can use TagMaps with your own data source to plot tags from your own application on a map. Finally, you can use our World Explorer tag data for your own map-based application. Check here for more details.

One last note for the curious people out there - you can learn more about how the World Explorer data is created here, or on our TagMaps FAQs. Enjoy!

Finally, a word of thanks - to Flickr users - keep uploading all these geotagged images that help us extract such wonderful data.