Talking Tweets: Sentiment Analysis in Twitter
The birth of Web 2.0 supposed a breaking down of the barrier between the consumers and producers of information. In other words, the web changed from a static container of information into an active element in which any user, in a very simple manner, could publish any type of information. The information generated by users varies widely; from publications in blogs or other fora to simple commentaries on their state of mind in social networks.
Nowadays, people usually publish several aspects of their lives on the Internet. At first, blogs and fora were the sites in which users posted their feelings, daily activities and thoughts on any topic which they considered interesting. Currently, users prefer to post their thoughts or daily activities on social networks. The most common types of posts are opinions about a product purchased, a marketing campaign, political activities carried out by the government or a service offered by the government. Furthermore, consumers usually seek the opinions of others before purchasing a product like a camera or a tourism service. Thus, several websites are available on the Internet which specialise in the publication of users opinions. Examples of these websites include: Epinions, Rotten Tomatoes, Google Shopping or the review section of tourism websites such as Booking.com.
The large amount of reviews that are constantly being posted on the Internet and the growing need to know other people´s opinions has piqued the interest of the Natural Language Processing research community leading to studies of the opinions of users. The discipline that focuses on the analysis of opinions is called Sentiment Analysis or Opinion Mining, which is defined as the computational treatment of opinions, feelings and subjectivity in texts.
Social networks have become a continuous flow of information. A clear example is the microblogging platform Twitter. Twitter users publish all kinds of information, disseminating views on many different topics: politics, business, economics and so on. Accordingly, Twitter has become a valuable source of information for the analysis of opinions. Therefore, the Sentiment Analysis community has focused on Twitter with the aim of obtaining opinions about any topic.
Since 2009, a lot of research papers have been published concerning Sentiment Analysis in Twitter, several methods have been presented dealing with the specific characteristics of these posts or other similar approaches while others intend to apply Sentiment Analysis techniques in a specific domain, and intend to present linguistic resources for use in the analysis of tweets.
Studies concerning Twitter are spreading and growing, so it is important to go over the already published work in order to learn about the research already carried out and continue along the path of resolving the issue of obtaining Twitter user opinions. This is the aim of the authors of the paper “Sentiment Analysis in Twitter” published in the journal Natural Language Engineering. The paper presents a report about the different computational methods applied to tweets, the different domains in which researches have attempted to obtain opinions, the linguistic resources available to the research community as well as workshops and congresses focused on Sentiment Analysis in Twitter.