AN IMPACT OF SOCIAL MEDIA VIA TWITTER ANALYTICS

Abstract

In today’s world micro blogging has become emerging connection medium for Internet users [1]. Many users share their views on different semblance in famous websites such as Facebook, Qzone, LinkedIn, Twitter and Tumblr. With increase in the user on social networking sites, many big giants and media organization are trying to achieve different ways to get these social media information so that they can know what people think about their quality, product and companies.

Many firms, Big Organization, Political parties as always keen in knowing if the people will sustain with their event, program or not. Many social NGO’s and Organization can ask people’s views on current topics, challenge for open debate etc. All such kind of information’s can be collected from such plenty of micro blogging websites. Here we represent a function which performs which will do classification based on tweets/retweets and calculate the impact of specific keyword/#Hashtag in Twitter. Currently twitter network is dazzled with huge no of tweets tweeted by its users. For an productive categorization and searing of tweets, user need to use suitable meaningful sentence and hashtags in their tweets. Twitter has a huge number of users which may varies from Politian’s, Celebrities, Actors, company representatives, an even country president uses twitter to express their views on social platform. By this ways we can collect the all possible text posts of users from different organization, companies, interest groups and different social groups [2]. In this project I propose generic functions/recommendation method of the tweets of the individuals/popular personalities that tweet which will create an impact on user mind after the tweets. If we have different groups of users and these tweets, we can easily create our methods so that we can find out the top most familiars users and top most familiars tweets from collected data. Hashtags/Keywords are then used to select to select most familiar tweets and user and then we can assign them some ranking values/scores to them. In future I will explore a more on more types of different categories by after-peak value, before peak value, and during-peak value popularity. It will be arousing to propose different recommendation methods.