Twitter and Data Mining

 As a public affairs officer in the US Navy, I have been exposed to the reliability and speed of Twitter.  I have both used and relied on Twitter as a effective tool to communicate with media as well as to track the pulse of a story or idea.

Last year, the US State Department was very interested in the service and reliability of Twitter.  A mini revolution occurred in Iran after the Presidential election took place in June 2008.  Many Iranians who claimed that the election was rigged protested via Twitter. Twitter was the ideal medium because it was fast, could be employed and leveraged by almost anyone with a computer or cell phone and was difficult for the government to track or block.

During a revolution, crisis or emergency, Twitter really showed its astounding strength and amazing news delivery ability. I would like to see this medium to continue to grow and expand as a communications and news medium and to continue to serve the public to make an impact in people’s lives, especially during times of emergency or distress.   

                   Twitter is a pretty simple concept that uses API that gives RESTful access to the Twitter database and does not have a lot of components or moving parts.  The simplicity of the user interface has resulted in the development of several hundred applications that has further fueled its growth.  Although these applications have boosted Twitter’s demand and exposure, Twitter does not see revenue from any of these applications.  Many of these applications are popular with users because of ease of use and their ability to provide a very specific but valuable function for free.

My goal is to focus on Twitter’s Revenue Stream.  That is because Twitter needs to find a long-term way of making money in order to stay viable. Although advertising is the primary way for other social media companies like Google and Facebook to make money, this technique may not be ideal for a micro blogging company like Twitter.  One possible option is to sell promoted Tweets to companies which will show up in the stream of Twitter posts, based on how relevancy.

                    Eliminate Problem Areas: Twitter unlike other social media sites like Facebook and websites like Google face challenges leveraging on advertising.  That’s because a tweet is limited to 140 characters and there is not much space to employ ads and to charge companies for them.  Twitter has recently (Oct 2009) worked a deal with Google and Bing to open up their tweets to these search engines. 

6                  Data mining is critical because no other tool provides the depth and breadth of information that Twitter can.  Twitter is not only good for marketers and advertisers, governments and analysts can use them to get a pulse on sentiment and behavior even before it is noticed or confirmed.

     

T                  Twitter can provide publishers and developers a way to plug into their data feeds to extract data and patterns. The advantage is that roughly 20% of tweets include branding.  The potential drawback is that you cannot verify the accuracy of tweets or even verify who the sender is.  Governments and large organizations can go on Twitter to send propaganda (such as Iran)

Since Twitter maintains and owns the tweets made by their members and they receive several million tweets a day all over the world, Twitter can really capitalize on their unique but value-added data flow and trend analysis obtained by data mining their information.  In Novemeber 2009, www.royal.pingdom.com conducted a study to measure how many tweets were sent per hour.  The highest number of tweets measured was 1,841,289 per hour on Oct 27, 2009.  At this rate, Twitter will process almost 10 billion tweets per year. 

Some of the data mining can be conducted simply by using the search function within Twitter. The vast majority of the data mining can be conducted using several hundred of the different applications that have been developed for Twitter.