•
It's hard to keep up with all the newly released movies and music these days, but a lightweight social network with a whole lot of smarts under the hood says it can now offer you personalized recommendations of new releases that suit your very particular interests. GetGlue is a semantic web browser plug-in that has, for years, been smart enough to recognize when you're looking at the same musical group across different websites, be that on Last.fm, MySpace or elsewhere. The service recently added a stream of recommendations of music, movies, books, magazines, wikipedia articles and other things you might like. How can it tell what you'll like when something is brand new, though? Today the service has launched a "new releases" section, where human editors rush to classify brand-new media. Then the semantic robots can serve it up to the right users, still hot out of the oven. It's pretty cool. Sponsor GetGlue founder Alex Iskold says he's learned a lesson similar to what formerly automated tech news aggregator Techmeme has learned: algorithms and user generated content can take you a long way, but there comes a point when it's good to hire some dedicated editors. The service asks you to like or unlike a wide variety of things. It then uses that feedback to build a taste profile to compare against things it finds put into its database and find the stuff it thinks you'll like. That's harder with new releases, though. "When something new is coming out, we don't know what it's like, so you need to have proffessionals tag it," Iskold told us. "We have two editors on staff who look across the spectrum of new releases each week. They draw the similarities between things in a deep way - the tagging system we use will be unvieled later. We use really eclectic tags to characterize what kind of zombie or vampire movie something is. We also use tags brought in from other systems and our users find cool new things really fast." The end result is a nicely displayed stream of big icons for personally recommended newly released movies, music and books. You think you're hip to your scene now? Wait until you've got a network of contacts, a semantic robot and real human editors all working together to bring you the freshest content in your weird little niche. To be honest, I've been testing it out today by switching from new album recommendations on Glue over to Apple's Lala.com , where it's easy to listen to full albums once for free. That's not the way Glue wants you to use it, but that's the way I like to use it so far. The Down Side It's an incredible system, when it works. GetGlue knows though that there are some challenges in this kind of game though. First, it's not easy to present this kind of flow of data to users without either overwhelming them or boring them. Many of GetGlue's latest changes are focused on making the user experience more pleasant: bigger images, collapsed bundles of shared items, etc. Can the service find a balance between giving you strong-enough recommendations on one hand and regularly offering up new recommendations on the other? In past versions of the product, I've received too few recommendations to keep me coming back. Hopefully new releases will scratch that itch. Iskold also says that after "liking" only 15 musical artists, I'm actually much less active than most of the 400,000 registered users of the service. Personally, I'm more drawn to the Wikipedia recommendations on GetGlue than anything else. The new releases in music might be roughly in the same sub-genres I usually listen to, but that doesn't mean they are any good. Finally, all this "liking" obviously begs the Facebook question. Writing as an ostensible Facebook competitor about that giant network's radical innovations unveiled last week, Iskold wrote the following in a widely-read article here at ReadWriteWeb about Facebook's Open Graph: "Time will tell where we land, but my gut is that positive things will come out of this. If nothing else, let's give Facebook credit for innovation and re-imagination the Web." Today he emphasized in speaking with me that Facebook is new to what it's just begun to do, but his company has been doing it for years. There's no guarantee that Facebook will get it right, he said. It's hard to say for sure that GetGlue has got it right, either. But as a work in progress, it's pretty darned good and today's new additions are very interesting. Discuss
Go here to read the rest:
GetGlue Adds New Releases to Recommendations Made by Human & Machine
Tags:
apple,
facebook,
kind,
like-or-unlike,
new-releases,
overwhelming,
really-eclectic,
recommendation engines,
semantic,
similarities,
tagging,
things-it-finds,
user,
Wikipedia
•
I love to sit on the beach. One of the coolest things about the beach is the number of layers of visual depth. Look at the sand and it's beautiful, but zoom your eyes in closer and you'll see a whole layer of life running around on the sand that you didn't see before. Look even closer and you can see individual grains of sand, water and light dancing between them. Look closer still and you see that each grain of sand is a unique object with its own texture. If your eyes are strong enough, or you have a machine to help you, you can see even more layers by looking closer still. That's what Twitter is going to be like with the launch of Twitter Annotations this Summer. It's a beautiful vision, with huge potential, but there's another way to look at this analogy: you don't build on the beach sand because it shifts too much. Will Annotations live up to its incredible promise? Sponsor What Annotations Are Last week Twitter announced a forthcoming feature called Twitter Annotations: it's a system for almost any metadata to be connected to any Twitter message when it's published. Inside every Tweet is now a space where you could put or find anything, including links out to further instructions or larger bodies of information. That's always been the case with the 140 characters of content - but now we're talking about systematic metadata intended for machines, to augment the content. The idea is dripping with potential, but also some risk. Isn't much of life's meaning found in the play between limits and the infinite? Twitter has been considering adding Annotations for at least two years, according to Platform Team member Raffi Krikorian. That's a relatively large portion of the company's young life. Every time a new bit of metadata was added to Tweets, like geolocation information was last Fall, the company would ask itself "should we be doing this, or should we just open up the platform for and and all metadata?" Now the company has decided to do just that. Twitter publishing tools can now add a description to any tweet their users publish, not as a part of the 140 character message, but as a small machine-readable metadata field that travels along with the content. What might this look like? We could see Annotations fields like: Link to a media file, like podcast enclosures, photos linked to, etc. Context about the Tweet like where was the author when it was published, maybe what the weather was like there at the time. Your Twitter publishing interface could offer you a special option to write reviews of movies, books, or links you're sharing. The ISBN of the book, a link to a preview of the movie and the number of stars in your rating could be included in the Tweet Annotations. Any way you can classify, describe, append or otherwise enrich a Tweet with words or numbers can be included in Annotations. You Tweet, you attach a characteristic or quality, you define the characteristic and then you provide a value of how or what that Tweet did relative to the quality being referenced. Twitter clients like Seesmic, Tweetdeck and more will make it easy for users to add these annotations. Yes, this is meaningful in large part because of the 140 character limit on Twitter messages themselves, but isn't much of life's meaning found in the play between limits and the infinite? From Annotations Come Analysis Annotating a single Tweet is uninteresting, it's when you hit the Twitter databases and gather together all the Tweets that share a characteristic that thinks get exciting. When those selected Tweets can then be cross-referenced with other sets of data from outside Twitter - that's when the word fecund starts feeling inadequate. Show me all the Tweets from my friends that have links to music and play me those songs. Twitter clients like Seesmic, Tweetdeck and others are going to make viewing that kind of data a whole lot easier. Tweetmeme's Nick Halstead believes that Annotations will be used most extensively to communicate webhooks, links to instructions for a Twitter client to follow. He thinks it will enable game play and help Twitter start acquiring more users again. "Because of the size of the data you can put in the annotations, I think people will come up with links to offsite resources. Seesmic is building their own platform for Windows to support plug-ins, but this reaches much further, but this lets Twitter clients augment a tweet with other services. Sf you were Stocktweets, you could attach a link in the namespace that's in stocktweets, Seesmic could follow that link back to Stocktweets and ask it how to render it. So you could put a chart and any other associated information. It's like FBML [Facebook Markup Language], the ability to embed applications inside the Twitter clients. Maybe threaded conversations. A game of Scrabble where the link points at a currently rendered scrabble board, so other people could look at the board and join in playing it. Annotations and webhooks would allow gaming to start happening on Twitter." Halstead believes an Alpha version of Annotations could be made available to developers in a month. How about showing me all the Tweets from anyone that are referencing the President of the United States (subject: POTUS?), analyze the sentiment in the messages, show me where those Twitter users were located and tell me how those local sentiments change over time. Send me an alert when one of those starts to shift radically. Show me all the Tweets by people in their 20's and in their 50's (imagine an author age tag in Annotations, why not?), living near the site of a disastrous event. How do those discussions differ? There are all kinds of interesting questions that could be tackled when the developer world's imagination runs wild on the terms of description applied to our messages. Of course it will be tempting to draw all kinds of conclusions from this rich data. We'll surely be able to draw a whole lot of value from it. "You can learn something from almost anything," Big Data cruncher and 80Legs CEO Shion Deysarkar says. "Just give me enough data, I'll figure out something." But let's keep in mind the words of social network scientist danah boyd, who wrote the following on her blog this morning: Time and time again, I see computational scientists mistake behavioral traces for cultural logic...Big Data creates tremendous opportunities for those who know how to assess the context of the data and ask the right questions into it. But mucking with Big Data alone is not research. And seeing patterns in Big Data is not the same as hypothesis testing. Patterns invite more questions than they answer. Tweet Power Politics Twitter's Krikorian says the site will probably list "trending annotations" just like it lists trending topics today. There will probably be a wiki where anyone can find out what namespaces are being used for what purposes. Really though, the classification system is going to be determined by the market. That's something that worries a lot of people. "People who believe in building standards are conerned about our blase attitude about how we want to run annotations," Krikorian says. He believes that the developer community will work things out for itself, just as it has in the past. "There has been a lot of emergent behavior around how to relate to tweets anyway, without our imposing much structure around it. The Twitter platform is continuously evolving - the developers will figure it out. Twitter developers iterate in public." That's likely to be cold comfort for people focused on the power of structured data standards. Many people are calling for Twitter to embrace the well-built efforts of the Semantic Web community. Krikorian says that 90% of Twitter developers don't know what the Semantic Web is but that there's certainly room for standards lovers to work within the Annotations scheme. Still, the absence of standard terminology could really be a problem. Annotations can't be changed retroactively, either. Krikorian says that major players will dominate the obvious use cases for Annotations and the company will monitor and highlight really innovative Annotations developed by people on the margins. We'll see how well that will work. Imagination will make the sky the limit for this publishing platform used easily by more than 100 million people around the world. But a shortage of forethought, planning and agreed-upon standards may bring that platform's aspirations back down to earth quickly in the future. Time will tell. Discuss
More here:
What Twitter Annotations Mean
Tags:
analysis,
Annotations,
data,
forthcoming feature,
grains of sand,
movie,
Nick Halstead,
people,
platform team,
power,
semantic,
summer,
tweet,
Twitter,
United States,
windows,
words
•
In our recent posts about Structured Data , we've emphasized that most of the current initiatives have been around uploading new data to the Web - whatever the format. The U.S. and U.K. governments have led the way with their 'open data' websites, but much of that data isn't 'linked' yet . In other words, it's online - but siloed. So how do we get to the next stage of the Semantic Web, linking disparate data sets together so that people can begin to use that data? The tipping point for the long-awaited Semantic Web may be when you can query a set of data about someone not too famous, and get a long list of structured results in return. I've decided to term this 'The Modigliani Test.' Sponsor Amedeo Modigliani is one of my favorite artists. He was moderately famous during the early 20th century and has something of a cult following nowadays. But he's not Da Vinci or Picasso famous. What I'd like to do in a Semantic Web is type the following query into a search engine and get back a large list of results: tell me the locations of all the original paintings of Modigliani. As of today, there's no place to type that query in and get a list of structured data . The closest I can find to doing that is the Artcyclopedia entry for Modigliani, which has a list of locations for Modigliani artworks. It's great that they have the location data listed on one web page. However it's not structured data, so we can't query it. There's also not much order to the data, we have no idea if this is a comprehensive list, it's not verified data, and so on. In summary, there's a lot of data on the Web about the location of original art works - but much of it is in traditional 'document' web pages. What we're after is a giant database of art works, which anybody can query and re-use. Here's an early, overly geeky view at what a Linked Data of painting locations would look like (hat-tip @dakoller ): The above is a far from comprehensive list of art works by Hieronymus Bosch (a search for Modigliani, by the way, brought up zero results). Plus of course we need a much more intuitive UI, so that non-geeks can use it too. What do you think, when will The Modigliani Test be passed on the Web? Discuss
See the original post:
The Modigliani Test: The Semantic Web's Tipping Point
Tags:
amedeo modigliani,
following-query,
Hieronymus Bosch,
linked-data,
list,
location,
locations,
modigliani,
original art works,
search-engine,
semantic,
structured data,
U.K.,
U.S.,
web,
words
•
Recently we looked at the state of Linked Data in 2010 , noting developments such as governments putting public data online and Thomson Reuters putting structure around commercial data using OpenCalais . In a follow-up post, we explained the distinction between Linked Data, Open Data and the Semantic Web . Georgi Kobilarov, who runs a Linked Data startup from Germany called Uberblic Labs , recently issued an interesting challenge on his blog. He asked: if we had a Web of Data, what would you build? Not to steal Georgi's thunder, but we think this is a great question to put to ReadWriteWeb readers too. Sponsor Here's Georgi's idea: "If we had a Web of Data, I would built an application for painless travel planning. It would integrate flight plans, train timetables, bus routes, car rental offers, etc. And the user would be able to just say: I want to go from A to B: Find me the best/cheapest/fastest routes. With a Web of Data, an application could do all that combining for me, the same way flight booking sites do that today for just flights." Here's my idea for an app that uses the Web of Data. I'd like a web site or app that allows me to discover the locations of original art works by my favorite artists, and then create travel itineraries for me to see some or all of those art works (most famous artists have their art works scattered around the world, in various museums and galleries). It's possible that there is a web directory of artists somewhere that has some or even all of this data already, but if so I haven't found it. I ask for this because every now and then I search the Web for a painting that I saw in a book. A recent example was a Modigliani painting that I was attempting to create a copy of , for my beginners acrylic painting class. The original painting was called "Portrait of Madame Hanka Zborowska." One of the results from Google told me that the original painting is located at the National Gallery of Modern Art, Rome, Italy. I could potentially spend hours hunting down the locations of Modigliani's paintings, using Google - and it's likely that some of the data isn't currently online. So it would be great if I could query one web site or app: tell me where all the originals of Modigliani's paintings are in the world, and draw me an itinerary for visiting all or some of them. Heck, maybe even book my flights and hotels! That's my example of what I'd build from a Web of Data. Now tell us what site or app you would like built , if the data was available on the Web. Discuss
Follow this link:
Open Thread: What Would You Build With a Web of Data?
Tags:
data,
distinction,
modigliani,
national,
open thread,
semantic,
user,
World
•
Yesterday we summarized some of the main developments in the Linked Data world over the past year. Linked Data is a W3C-backed movement that is all about connecting data sets across the Web. It can be viewed as a subset of the wider Semantic Web movement, which is about adding meaning to the Web. However, there is some confusion in the Semantic Web community about the crossover . To add to the confusion, there is a term called 'Open Data' that is being bandied around too. This commonly describes data that has been uploaded to the Web and is accessible to all, but isn't necessarily "linked" to other data sets. So what's the beef with all of these terms? In this post we seek clarity! Sponsor The Difference Between Open Data and Linked Data In the discussion over yesterday's post, a few people tweeted that the U.K. government's public data website Data.gov.uk is mostly populated with 'Open Data' and not 'Linked Data.' But what does that mean? It means that much of the data on the site is available to the public, but it doesn't link to other data sources on the Web. It could be data that has been uploaded in CSV format (i.e. spreadsheet data), which Sir Tim Berners-Lee said in an interview with me last year is a common occurrence with government departments. Or it could be data in another non-Web format. Screen from a Tim Berners-Lee presentation on Linked Data , circa 2008 Titti Cimmino put it nicely : Open Data is simply 'data on the web,' whereas Linked Data is a 'web of data.' However, the idea of Open Data is to turn it into Linked Data. As John S. Erickson pointed out , the first priority of Data.gov.uk (and its U.S. counterpart) is to publish lots of Open Data. The next step is to work towards linking it all up. This is already starting to happen. Answering a question I posed on Twitter, Kingsley Idehen confirmed that Data.gov.uk is currently a combination of Open Data and Linked Data. Linked Data and The Semantic Web So may we then suggest that the idea of Linked Data is to turn it into a Semantic Web? Or are they the same thing already? Lorna Campbell from the University of Strathclyde in Scotland tackled those and other questions in an excellent post earlier this month. She started by warning of the potential for another "holy war" about terminology. I won't delve into that in this post, however this excerpt from Campbell's post gives you a flavor of the terminology angst: "Some argue that RDF is integral to Linked Data, other suggest that while it may be desirable, use of RDF is optional rather than mandatory. Some reserve the capitalized term Linked Data for data that is based on RDF and SPARQL, preferring lower case "linked data", or "linkable data", for data that uses other technologies." Even Wikipedia can't define Semantic Web... Campbell quotes from a number of other articles, in trying to come to a conclusion about how Linked Data and the Semantic Web relate. Perhaps the best definition she found was this one by Paul Walk : data can be open , while not being linked data can be linked , while not being open data which is both open and linked is increasingly viable the Semantic Web can only function with data which is both open and linked " Why This Matters So there you have it, Linked Data is NOT the same as the Semantic Web. It's also not necessarily open, in other words accessible to developers. Whatever the definitions, the key points about all of Open Data, Linked Data and the Semantic Web, are: data is being uploaded to the Web that wasn't online before (e.g. much of the data on Data.gov.uk). structure is being added to the data using Linked Data and/or Semantic Web technologies. The bottom line is that the more data we have on the Web that is linked and has defined meaning, the smarter our web applications will be. This is why these activities are so exciting, despite the terminology confusion! Image credit: Semantic Web Rubik's Cube, dullhunk Discuss
See the original post here:
It's All Semantics: Open Data, Linked Data & The Semantic Web
Tags:
Campbell,
data,
data-gov-uk,
John S. Erickson,
Kingsley Idehen,
Lee,
Linked,
linked-data,
Lorna Campbell,
Paul Walk,
Scotland,
semantic,
sir tim berners lee,
Sir Tim Berners-,
structured data,
terminology,
tim berners lee,
Tim Berners-,
U.K.,
U.S.,
university,
web,
Wikipedia,
words