The Twittering of the Search World
by David Mattison
Aloha Fact and Image Finders, Victoria, BC
In 4 short years, Twitter [http://www.twitter.com] has gone from being a fad to a phenomenon to an essential component of any social networking and online business suite.
It has also evolved into a major player in the search engine and information dissemination business. According to an SEO (search engine optimization) industry report from VentureBeat.com, Twitter co-founder Biz Stone announced on July 6, 2010, that daily queries at Twitter Search [http://search.twitter.com] had reached 800 million or (24 billion a month), up 33% since April 2010 (“Twitter Search Queries up 33 Percent From April to 800 Million per Day,” Kim-Mai Cutler, VentureBeat.com [http://social.venturebeat.com/2010/07/06/twitter-search-800-million-queries]).
Data about who’s using Twitter and how varies widely depending on its source. Presumably, the most reliable information comes from Twitter itself. In April 2010 at its Chirp developers’ conference, Twitter released these statistics: 105,779, 710 registered users; 180 unique visitors per month; 75% of traffic coming from outside twitter.com; 3 billion requests a day through its API; about 600 million search queries per day to Twitter Search; 55 million Tweets per day, of which 37% come by phone; and 300,000 new registered users per day (“Twitter User Statistics Revealed,” April 14, 2010 [http://www.huffingtonpost.com/2010/04/14/twitter-user-statistics-r_n_537992.html]). comScore, however, “estimated 8.6 million worldwide unique visitors to Twitter.com in April ….” (“Costolo: Twitter Now Has 190 Million Users Tweeting 65 Million Times a Day,” Erick Schonfeld, June 8, 2010 [http://techcrunch.com/2010/06/08/twitter-190-million-users]).
The Twitter Firehose, a live content feed of all public Tweets, has not only been accessible to specialized search engines that query only Twitter content but also to the three major search engines (Google, Microsoft Bing, and Yahoo! Search), selected other real-time search engines, and the Library of Congress as a one-time static archive (“How Tweet It Is!: Library Acquires Entire Twitter Archive,” Matt Raymond, Library of Congress blog, April 14, 2010 [http://blogs.loc.gov/loc/2010/04/how-tweet-it-is-library-acquires-entire-twitter-archive]).
In this article, I look at some of the ways in which the three biggest search engines, along with Twitter Search and several other third-party vendors, attempt to fine-tune the ceaseless cacophony of Twittering tweeps to extract useful information for your queries. I will cover search engines which draw their entire content from Twitter, along with a few other examples that pull in data from other microblogging sites, such as Identi.ca [http://identi.ca]. I’ll introduce you to some various Twitter-centric general and search widgets (web browser tools) and apps, and we’ll look at a few of the Twitter metric tools. Because of the large volume of public data Twitter continues to amass, academic and industry researchers, through arrangements with Twitter, have been able to advance their studies of data mining techniques using what is likely the world’s largest, publicly available mass of microcontent. I’ll point you to a few of those results.
Because I wrote this article during the FIFA World Cup soccer championship (June 11–July 11, 2010) and the Gulf of Mexico British Petroleum Deepwater Horizon oil drilling platform catastrophe also occurred, many of my examples utilize search terms such as world cup, vuvuzela, deepwater horizon, and oil spill. I did not, however, conduct any rigorous, scientific testing. As but one example of the extent to which the corporate world has fully engaged with Twitter, one of my world cup searches quickly came upon the AT&T-sponsored CupBuzz: World Cup 2010 [http://www.titletweets.com/cupbuzz]. I’m sure the name is intended to reflect the many Twitter conversations around this international soccer championship and not the mass sound made by the vuvuzela. AT&T, a sponsor of ESPN’s World Cup broadcast, also sponsored the associated CupBuzz Twitter feed [http://twitter.com/CupBuzz2010].
A Little Twitter History
Twitter, the company, was established in March 2006, and Twitter, the service, was publicly launched that July. Twitter was originally designed for the mobile phone market via text (SMS; short message service) messaging and then evolved into its own web subculture. Twitter is categorized as a microblogging or micro-messaging platform because you are limited to 140 characters per Tweet or message. This limit relates to the SMS standard of 160 characters with 20 characters subtracted for Twitter itself to use. (For details on the historic background for that limitation, read Mark Milian’s May 3, 2009, Los Angeles Times Technology blog report, “Why Text Messages Are Limited to 160 Characters” [http://latimesblogs.latimes.com/technology/2009/05/invented-text-messaging.html]).
The Twitter Help Center [http://support.twitter.com] contains much useful information, including the Twitter Glossary [http://support.twitter.com/articles/166337-the-twitter-glossary], through which you can learn the true meaning of the obtuse phrase “fail whale.” The history of this cute visual that functions as Twitter’s 503 (service unavailable) server error message is also recounted in Sarah Perez’ “The Story of the Fail Whale” (ReadWriteWeb, July 17, 2008, [http://www.readwriteweb.com/archives/the_story_of_the_fail_whale.php]).
Since I don’t use a mobile device to Tweet and my primary focus is general public accessibility to Twitter content and third-party services at no cost, my coverage of Twitter search and services is limited to openly available websites, some of which also offer a premium (user pay) service. Many of the free websites and applications do, however, require you to register for a free account or login with your Twitter account in order to make full use of their services. No doubt Searcher Library Mobile columnist Gerry McKiernan will cover the many ways in which librarians have incorporated Twitter Mobile [http://m.twitter.com] into their service offerings.
With more than 50,000 Twitter-fed applications out there (Twitter Blog, March 1, 2010 [http://blog.twitter.com/2010/03/enabling-rush-of-innovation.html]), no one can possibly try them all. You can locate many Twitter-based search services through the official Twitter Blog [http://blog.twitter.com], hosted by Google’s Blogspot service, or Twitter’s own Twitter feed [http://twitter.com/twitter], followed by 3,337,619 users (some of whom could be bots), far fewer by 2 million than are paying attention to Britney Spears, at least according to Twitaholic [http://twitaholic.com].
The Big Three
How do the big three general search engines handle Twitter content? Google and Bing have a partnership arrangement with Twitter, announced in late October 2009, that gives them access to what Twitter calls its full-force Firehose of public Tweets [http://blog.twitter.com/2009/10/google-nice.html; http://blog.twitter.com/2009/10/bing-goes-dynamite.html]. Yahoo! followed in February 2010.
Google [http://www.google.com] plays no favorites with Twitter and other social networking content; its default search results screen displays “Everything.” As of July 13, 2010, you can view all real-time (live) Tweets and other content Google indexes by either selecting the Updates link or the Latest link in the Google results filtering menu on the left. The only way to filter for Twitter.com content in the live feed is to add site:twitter.com before or after your search terms. In addition to normal search terms you can also search for Twitter hashtags such as #vuvuzela or #worldcup.
I discovered, however, that you can’t search for and retrieve results in the Updates or Latest displays for a tweep’s username, assuming you happen to know it. If you know a person or organization’s Twitter name, you’re better off entering it directly. No matter how carefully you construct a search for britneyspears site:twitter.com, @britneyspears site:twitter.com, or “twitter.com/@britneyspears”, you will see mentions of britneyspears by all Twitter users. If you don’t know the tweep’s name, at least the Google search might give it up.
Curiously, the Google Advanced Search does not offer a Latest date option; the closest you can get to Latest in the Google date option is the “past 24 hours.” Once you get those results, you can select Updates or Latest. The only interaction you have with the Google real-time content feed through the Latest display is to pause it if you spot something that interests you. I did not check to see what would happen if I left it on pause for too long and then resumed the live update. Would the new content scroll by quickly? Or would it jump ahead by seconds, minutes, or hours? I was surprised to see among the Google Latest results a link to the Microsoft Bing Social subsite, as well as various spamvertisements.
The Updates display gives you a timeline bar graph view of your results, but you cannot pause the display. The Updates timeline, with spikes representing high levels of Tweets on your search terms, only extended back to February 2010. Google introduced this service on April 14, 2010, and described it in its blog post, “Replay It: Google Search Across the Twitter Archive” [http://googleblog.blogspot.com/2010/04/replay-it-google-search-across-twitter.html]. Dylan Casey, product manager for Real-Time Search, promised that “soon you’ll be able to go back as far as the very first tweet on March 21, 2006,” but as of July 13, 2010, that functionality was not available.
Assuming it’s still around, you might take a look at another Google experimental offering called Google Trends [http://www.google.com/trends]. Google Trends is the company’s own analysis of searches and news content aggregation back to 2004. The main page includes a search box for the “Latest” (real-time) content. Not all the pages from this site are labeled as a product of Google Labs, so until and if it ever graduates to the real world, you could use it to query and graph results from up to five search words or phrases. You can also examine data for specific websites, though you will find no data from Google as compared to Bing and Yahoo!.
Microsoft Bing [http://www.bing.com] gives you direct access to Twitter and real-time Facebook updates through Bing Social [http://www.bing.com/social], which first launched in October 2009 as a Twitter-only search service [http://www.bing.com/twitter], redirects to Bing Social, and was rebranded when it added Facebook content on June 9, 2010 [http://www.bing.com/community/blogs/search/archive/2010/06/09/use-bing-social-to-search-facebook-and-twitter.aspx]. Some additional information on how Bing Social works appears in this Bing Community Blog entry [http://www.bing.com/community/blogs/search/archive/2010/04/13/get-the-latest-on-twitter-with-bing-social-search.aspx]. Bing Social lacks the Google News focus of Google Trends. Bing Social is also afflicted at times by spamvertisements that become trendy and then disappear. At one point on July 7, 2010, I noted that six of the 10 Hottest Social Topics were essentially spam.
Twitter announced in late February 2010 that Yahoo! was given access to the Twitter Firehose and would integrate these Tweets into “Yahoo! Search as well as other popular products and properties, including the Yahoo! Homepage, Yahoo! Mail, Yahoo! Sports, and more. Yahoo! will also be able to build unique Twitter clients into their properties making it easier for folks to tweet wherever they feel comfortable within the Yahoo! network” (“Expressing Great Joy Or Excitement,” Feb. 23, 2010 [http://blog.twitter.com/2010/02/expressing-great-joy-or-excitement.html]). Since Yahoo! is both a search engine and a content portal, you will find, as Twitter promised, more Twitter content appearing in different parts of Yahoo!, along with content from other social media/networking destinations such as Flickr. So far, at least from what I can determine from the often confusing mishmash of Yahoo! subsites and changing layouts, Twitter content is only directly accessible as of July 7, 2010, through Yahoo! Search with Twitter appearing as a filter label. If you click the link to view more Twitter content, however, you’re taken to Twitter Search [http://search.twitter.com]. If you have a Yahoo! account, you can use Yahoo!’s own social networking and content aggregation service, Yahoo! Pulse [http://pulse.yahoo.com], to set up a feed for Twitter content. By connecting with other Yahoo! users who also Tweet, you can view their Twitter content. Yahoo! Pulse also provides you with access to a large number of Yahoo! content sources for your Pulse Home and Profile sections.
The Basics of Twitter Search
Twitter Search [http://search.twitter.com] made its first public appearance in late April 2009. It is accessible from your Twitter profile page. It serves as Twitter’s own search engine of its users’ public Tweets. Twitter Search was based on the work of Summize.com, which Twitter acquired in July 2008 (“Twitter Confirms Summize Acquisition,” Jennifer Leggio, July 15, 2008 [http://www.zdnet.com/blog/feeds/twitter-confirms-summize-acquisition/150]). The Twitter blog announcement on its new search engine noted:
As public tweets fly in from around the globe, we analyze them to detect when certain words or phrases occur with higher frequency. These trending phrases are surfaced in the Twitter home page just under the new search box and they’re updated throughout the day. Built on our search technology, trends are a compelling if rudimentary way to explore a collective global consciousness.
Assuming you have a Twitter account, with Twitter Search, “You can ask Twitter to build you a fresh timeline of updates based on a keyword or phrase,” and you can create a list of saved searches. No advanced search functionality is available from your Twitter profile (user) page. You need to go to Twitter Search itself to access the Advanced Search. There you can see some of the unique aspects of Twitter addressed through its own search form: hashtag (prefixed by #, any single stream of characters up to a space) searches, Tweets between individuals or which refer to another individual, location information between 1 to 1,000 miles or kilometers, “attitudes” equated to the emoticon symbols of :) (positive) or :( (negative) or an interrogative symbol (?), Tweets with URLs, and retweets.
What’s Better and Different Than Twitter Search?
The only reason that Twitter-based applications are a growth industry is because of the Twitter APIs (application programming interfaces) that provides developers with access to all Twitter data marked by users as public. The basic APIs, including the Search API, are free services, but as with many things in life that are free, there are limits on what developers can do and how they can accomplish their automated tasks without running afoul of the Twitter API police. Some developers have proven their innovative worth to Twitter, such as Ellerdale [http://www.ellerdale.com]; Collecta [http://collecta.com]; Kosmix [http://www.kosmix.com]; Scoopler [http://www.scoopler.com]; twazzup [http://www.twazzup.com]; CrowdEye [http://www.crowdeye.com]; and Chainn [http://www.chainn.com], a Facebook application developer.
In late February 2010, some developers received access to the full Twitter Firehose through the Streaming API [http://blog.twitter.com/2010/03/enabling-rush-of-innovation.html], a licensed use for which Twitter receives revenues (“What Is Taking a Sip From the Twitter Firehose Going to Cost?” Liz Gannes, Gigaom, March 1, 2010 [http://gigaom.com/2010/03/01/what-is-taking-a-sip-from-the-twitter-firehose-going-to-cost-you]). In a July 2, 2010, report Gannes noted that, according to a Twitter spokesperson, the number of companies with access to the Firehose is 15 to 20 [http://gigaom.com/2010/07/02/whats-on-deck-for-twitters-platform-app-promotion-and-another-dev-conference]. You can find further information and full documentation about the API on creating applications that work with Twitter data or about integrating Twitter content into your own site through Twitter Developers [http://dev.twitter.com].
I’ve tried to categorize a selection of the numerous Twitter search engines by their different methodologies or search facets that help them stand out as worthy of your attention. The significance of why these search engines have proliferated is underscored by two facts — the sheer volume of public data passed to and then gushing out of the Twitter Firehose and, according to the Twitter API documentation, “Tweet text can potentially mention other users, lists, contain URLs, and contain hashtags — in fact, something like 50% of tweets contain at least one of those” [http://dev.twitter.com/pages/tweet_entities].
General Twitter Search Engines and Portals
While not all the listed Twitter search engines and portals may offer a distinctive search experience, I’ve included a few because of their ease of use and/or extra service offerings.
I admired the beta Topsy [http://topsy.com] for its simple, clean search interface, clearly modeled on Google. Topsy describes itself as “a search engine that ranks links, photos and tweets by the number and quality of retweets they receive. The quality of retweets is determined by the influence of the Twitter users” [http://code.google.com/p/otterapi]. Topsy’s Advanced Search will not overwhelm you with choices, but, if you need more powerful query options, look at the Twitter search operators screen [http://topsy.com/operators]. You can filter your Tweets search results in a timeline by hour, day, week, or month, and sort your results by their relevance or currency (timeline). Topsy’s representation of Experts, which it tags as either Highly Influential or Influential, is calculated “using all historical retweets: millions of real, public statements indicating who’s listening to whom” [http://labs.topsy.com/influence]. Topsy also lists Trending results ranked between the Top 100 and Top 20,000 Tweets; you can also filter or refine these results by adding additional search terms. You can log in to Twitter from Topsy, grab Topsy buttons for your website and a bookmarklet for your browser, and developers can tap into Topsy through its Otter API.
Calling itself a “Twitter Search Dashboard,” TweetGrid [http://tweetgrid.com] fulfills the wishes of every search addict by accommodating up to nine different Twitter feed searches through which, when logged in to your account, you can also interact with what you’re reading and post to your own account. There’s even a handy hashtag box if you want to add one to your Tweet. A TweetGrid Twitter Search Widget is also available. Chad Etzel, the developer of TweetGrid, also sells a Twitter Search data-push service called TweetHook [https://tweethook.com] that uses a WebHook [http://webhooks.pbworks.com] to send Tweets based on search queries to a URL.
Sort of a lesser version of TweetGrid, Monitter [http://www.monitter.com] lets you enter up to three keywords to follow up to three Twitter feeds at once in either a dark or light screen display format. You can narrow the location of the content to within 10 to 100 kilometers or miles of a location. Monitter defaults to a “trends” display based on whatever the current set of hot Tweets are at the moment, and you can add a Monitter widget to your own web server.
Similar in concept to Monitter, I enjoyed the visual appeal of Twitterfall [http://twitterfall.com] for its elegant, colorful, vertical, screenwide, rolling display of multiple Twitter searches, along with its many query options and display settings. As with other services, by logging in through your Twitter account, you can interact with the Tweets.
Pretty in pink Tweefind [http://www.tweefind.com] lets you filter Tweets by only the English language and only those containing links. You can link to YouTube videos relating to your search, other websites and blogs, view a TweetMeme list of the most retweeted links, or link to Google or Bing search results for your query. Hover your cursor over any of the Tweets and a series of Twitter popup interactions appear, but you have to sign in to your Twitter account to use them. Tweefind also has its own API and website widgets.
Launched in 2007 by SterryIT, TweetScan [http://www.tweetscan] tracks popular searches, presumably through its site, provides an email alert service, a Twitter user search function, and a Twitter account backup.
Anyone who’s blogged for several years would have encountered Technorati [http://technorati.com], “the first blog search engine” [http://technorati.com/about-technorati]. Based on Sawhorse Media’s MuckRack platform, Twittorati [http://twittorati.com] meshes Technorati content with Twitter’s. Twittorati, however, is the perfect example of why authority or influence does not always equate to real time. When I visited on July 16, 2010, none of the Latest Twittorati Chatter was less than 3 hours old, and one of the Twittorati topic categories, Technology, had no Tweets at all.
Geolocation and Other Specialized Twitter Search Engines
With Twitter’s integration in mid-June 2010 of two popular geolocation services (foursquare and Gowalla) as part of a new Twitter feature called Twitter Places (an enhanced version of the existing opt-in Tweet with Your Location feature [http://blog.twitter.com/2010/06/twitter-places-more-context-for-your.html]), I expect we will see more applications devoted to tracking Tweets and tweeps by their supposed location. This can have its downside, as PleaseRobMe.com made clear at the beginning of 2010. The Twitter Blog announcement confirmed that “We are releasing API functionality that lets developers integrate Twitter Places into their applications.”
Twellow [http://www.twellow.com], the “Twitter Yellow Pages,” offers a map-based service called TwellowHood through which you can find other tweeps to connect with. In North America alone, TwellowHood has data for close to 3 million Twitter users. Twellow has an extremely well-developed classification system through which a Twitter account holder can, upon registering with Twellow, add him- or herself to up to 10 categories. Once you’ve done this, the system will suggest other users for you to follow. Twellow statistics are staggering: 3.69 billion followers, 23.7 million Twitter profiles (as of July 15, 2010).
Some Twitter search engines look for references or links to specialized content such as Twitmatic [http://www.twitmatic.com], “a real-time stream of videos being shared on Twitter.”
All the News That’s Fit to Tweet
While much of what’s called “news” coming through Tweets is nothing but rumor, gossip, innuendo, and people retweeting what they’ve read, seen, or heard through traditional media outlets, Twitter has proven its worth as a real-time citizen journalism application and as a traditional reporting tool.
Launched in April 2009, twazzup [http://www.twazzup.com] bills itself as “a leading real-time news platform.”
DailyRT, which once featured only popular Tweets, was rebranded as Chirps [http://chirrps.com] and targets “news” content from tweeps.
Tracking Trends With Tweets
Just about every Twitter search engine I looked at offered its own interpretation of what constitutes a trend. In general terms, trends are based on an interpretation of searches by all users of the Twitter service, not just those who have a Twitter account. In Twitter chief scientist Abdur Chowdhury’s Twitter blog entry, “Top Twitter Trends of 2009” [http://blog.twitter.com/2009/12/top-twitter-trends-of-2009.html], he notes: “Among all the keywords, hashtags, and phrases that proliferated throughout the year, one topic surfaced repeatedly. Twitter users found the Iranian elections the most engaging topic of the year.” The Twitter Fan wiki page on Trends Apps [http://twitter.pbworks.com/Trends-Apps] will give you some idea of the kinds of trend information tools developers have created.
The brainchild of Martin Dudek, his Twopular [http://twopular.com] Tweet trend tracking and trend search tool presents a range of reporting intervals from 2 hours prior to your visit, 8 hours past, a day, a week, a month, or everything back to Dec. 5, 2008, when the service launched. Twopular also features a tag cloud generated through Thomson Reuters’ free OpenCalais web service [http://www.opencalais.com]. Other trend tag visualization experiments here include sparklines and bar graphs of trends and a trends heat map.
Looking for trends in Twitter content through its access to the Twitter Firehose is the goal of the Ellerdale Project [http://www.ellerdale.com] and the alpha version of its Twitter trend tracking tool [http://trends.ellerdale.com], which began in 2008 as a semantic search engine (“Video: The Ellerdale Project Makes Sense of Twitter,” Liz Gannes, Gigao [http://gigaom.com/2010/07/06/video-the-ellerdale-project-makes-sense-of-twitter]).
Authority, Influence, and Popularity Counts Among Tweeps
While primarily focused on tracking what’s popular on Twitter, you might also consider using TweetMeme [http://tweetmeme.com] as a finely crafted web-based Tweet reader. TweetMeme was launched in January 2008 by Favorit Ltd in England. Dozens of “channels” and content categories are available. For each Tweet, you’ll see a popularity time stamp such as Made Popular 55 mins ago. TweetMeme includes photo- and video-based content as well. You can integrate TweetMeme through plug-ins and buttons with the three major blogging platforms (WordPress, Typepad, and Blogger) and several content management systems. TweetMeme also has its own API. By logging in with your Twitter identity, you can also access My TweetMeme [http://my.tweetmeme.com], which gives you a trial TweetMeme Analytics and some additional TweetMeme customization. If you like monitoring multiple, live Tweet feeds, you may like TweetMeme’s TweetTabs [http://tweettabs.com], which uses the Tweet Search API.
Rather than rely on an automated approach to defining influence and authority, Sawhorse Media [http://sawhorsemedia.com] took a more direct approach: Find all the tweeps associated with a particular topic, discipline, or activity and create a Tweets portal around them. Some examples are MuckRack [http://muckrack.com], launched in 2009, of English language journalists; Science Pond [http://sciencepond.com] of scientists; and the hilarious Petfeed [http://thepetfeed.com], “pets on Twitter.” (Charles Darwin would be proud.)
With a bit of sociability edginess thrown in through its CrowdEye Rank that gauges a person’s Twitter influence, CrowdEye [http://www.crowdeye.com] lets you follow people by indexing everything that comes out of the Twitter Firehose. You can also filter results by location and sort them by time or relevance. If you sign in with your Twitter credentials, you’ll also get Personalized Follow Suggestions.
Created by Damon Cortesi, the beta TweepSearch [http://tweepsearch.com] “By default … searches the name, screen name, bio and location fields of Twitter profiles.” You can also search by a tweep’s location and use various common search operators or delimiters. When I visited on July 15, 2010, the site claimed to have indexed “over 13 million Twitter profiles.” Cortesi even has his own TweepSearch API.
I was unimpressed by the effort of Twitority [http://www.twitority.com] at gauging the authority of Tweets on any given query. Once you have a set of results, which defaults to the useless “any authority,” you can filter them by “a little authority” or “a lot of authority.” All the searches I tried suggested to me that either the most authoritative sources can’t be bothered to Tweet or there’s something seriously wrong with Twitority’s algorithm. The similarly named Twithority [http://twithority.com] gave no explanation of how results were derived for its “twitter search by – authority.” It appeared to base its “most recent” and “top rank” results on the number of followers for a given Twitter user. According to a March 5, 2009, Information Today, Inc. Weekly News Digest, Twithority is owned by the Canadian company, Tsavo Media. Paula J. Hane wrote that Twithority “sequences results by rank (with highest ranking users first) and time (with most recent Tweets first), within the top 10,000 Twitter users” [http://newsbreaks.infotoday.com/Digest/Tsavo-Launches-Twithority-Search-Engine-for-Twitter-52897.asp].
Searching for Sweet and Sour Tweets
Launched in June 2009, TipTop [http://feeltiptop.com] calls itself the Insight Engine and only uses Twitter content, drawing upon what the Twitter Advanced Search calls “attitudes” in an attempt to rate Tweets as either Tips (positive), piTs (negative), or Neutral. You can try searching for your own tips, though you have to construct your query carefully to generate useful results. The search box generates a continuous list of either a “TipTop cool query” (TipTop icon) or a “Twitter trend query” (a blue bird icon). As with many other Twitter search engines, if you have a Twitter account, you can interact in various ways with the Tweets.
You’ll find a much simpler example and display of attitude measuring at TweetFeel [http://www.tweetfeel.com], which bases its trends analysis on searches for positive or negative “feelings” within a Tweet.
Hashing the Night Away
Hashtags are Twitter’s answer to creating some meaningful content out of a specific series of characters entered by a user and preceded by the # symbol. Everything following the # sign up to a space character is part of the hashtag. Some tweeps don’t bother reading the extensive Twitter online help, as you can see from a search of #worldcup and world cup at hashtags.org [http://hashtags.org], a hashtag tracking and search service, where I saw many instances of #world cup resulting from the latter query.
Most of the Twitter search engines that offer advanced search capabilities can search for hashtags. Other specialized search engines whose main purpose is enhancing that experience in some fashion include Twemes [http://twemes.com] and for mobiles, http://m.twemes.com; TweetChat [http://tweetchat.com]; and HashParty [http://www.hashparty.com]. No, not that kind of celebration, but “the only twitter hashtag explorer that reveals the who behind the tweets!”
In Love With Links?
Looking for links posted in Tweets? True to its name, Twitt(url)y [http://twitturly.com] tracks and displays Tweets based on their URLs in four views: Everything, News, Pics, Videos. According to its About page, “Each time someone tweets a URL to their followers on Twitter, Twitturly takes note of it and applies it as a vote for that URL. The more votes a URL has in the last 24 hours, the higher it ranks on Twitturly’s Top100. … Whether people link directly to the final URL, use TinyURL, Snipurl (snurl), or any other URL shortening service, we always count the ‘votes’ correctly because our spiders actually visit every single site before it gets displayed here. … To keep things fresh and the quality high, the Twitturly Top100 only shows the 100 most popular URLs over the last 24 hours” [http://twitturly.com]. The Twitturly Twitter Profiles (user) search seemed somewhat squirrelly. It could not find my profile even though my Tweets are public, despite Twitturly suggesting they were protected.
BackTweets [http://backtweets.com] offers a free service, with a premium version with more features under development, that lets you search for any URL directly, including by date ranges through the advanced search, and see any Tweets that have linked to that URL. You can also look for Tweets containing links associated with your search words.
Variations on Visualizations of Tweets
Canadian data miner and visualizationeer Jeff Clark’s Neoformix blog [http://neoformix.com] is well worth exploring. Examples of his stunning work with Twitter content include Twitter Spectrum [http://www.neoformix.com/Projects/TwitterSpectrum/TwitterSpectrum.html], Twitter Venn [http://www.neoformix.com/Projects/TwitterVenn/view.php], or TwitArcs [http://www.neoformix.com/Projects/TwitArcs/TwitArcs.html], various streamgraphs, Tweet-searchable word clouds or maps in shapes based on the Tweet content such as the one of U.S. President Barack Obama [http://neoformix.com/2009/ObamaTwitterWordMap.html]. You can find many more visualization examples utilizing Tweets on his Projects page [http://neoformix.com/Projects/portfolio]. A Harvard Business Review blog profiled his work in “Four Ways of Looking at Twitter” by Scott Berinato (Feb. 18, 2010, [http://blogs.hbr.org/research/2010/02/visualizing-twitter.html]).
Through Twitzalyzer [http://twitalyzer.com], a free and premium analytics tool that looks at a user’s statistics, I found a visualization application called MentionMap [http://apps.asterisq.com/mentionmap], released on Oct. 14, 2009. Simply enter any user name to generate a node-type visualization graph based on its rate-limited access to the Twitter feed. You can embed MentionMap in your own website.
To find many other examples of Twitter data visualization, explore Information Aesthetics [http://infosthetics.com] and Nathan Yau’s FlowingData [http://flowingdata.com]. Yau also offers Your.FlowingData [http://your.flowingdata.com], which is data you might want to track over time and which is collected through direct messages as opposed to public Tweets. Since this is Yau’s personal project, you might want to think twice about the kind of personal data you input. Still, he seems to have accomplished in a simple, elegant way, though not through Tweets themselves, what Todd Fast and Jiri Kopsa proposed with Twitter Data [http://twitterdata.org] in May 2009.
Academicians coupled with private R&D entities have been studying, analyzing, and looking for ways to improve the Twitter search experience, as demonstrated by the work by Gene Golovchinsky and Miles Efron in their paper “Making Sense of Twitter Search” (CHI 2010, Workshop on Microblogging: What and How Can We Learn From It? April 11, 2010; PDF paper at http://www.cs.unc.edu/~julia/accepted-papers/chi2010-accepted.html). For more examples such as theirs see the Bibliography of Research on Twitter & Microblogging [http://www.danah.org/researchBibs/twitter.html] and the overlapping Bibliography of Research on Social Network Sites [http://www.danah.org/researchBibs/sns.html] maintained by Danah Boyd [http://www.danah.org and http://twitter.com/zephoria], a social media researcher and scholar with Microsoft Research New England and Harvard University.
Ever since it took flight, the future for Twitter has looked positively rosy. Twitter has its share of detractors and has weathered some controversies, most recently among the developer community. Gigaom’s Liz Gannes wrote of some new features that should help strengthen Twitter’s reputation and brand: “The Highly Anticipated Capability to Add Annotations to Tweets” [http://gigaom.com/2010/06/20/twitter-annotations-are-coming-what-do-they-mean-for-twitter-and-the-web/] is due in the second quarter. A new infrastructure for what’s called ‘user streams’ is behind schedule, but promises to make tweets and other actions get pushed out in real-time ….” (“What’s on Deck for Twitter’s Platform: App Promotion and Another Dev Conference,” July 2, 2010 [http://gigaom.com/2010/07/02/whats-on-deck-for-twitters-platform-app-promotion-and-another-dev-conference]). Finally, I anticipate that as Twitter continues to monetize and diversify its revenue sources, we can also expect to see service improvements (fewer fail whale screens) and many more kinds of Tweet search and analysis applications.