GPLSI Social Observer is an application that retrieves tweets from the Twitter social network on a specific topic and automatically assesses the opinions expressed in the messages. This allows you to track people’s opinions about a celebrity, politician, soccer team, product, stock market investment, or even an election. In this way, and thanks to this tool, it is possible to analyze and predict opinions and trends. This analysis can be represented both geographically on a map and with pie, line, and bar charts.
Trend analysis is a relatively recent research area of great interest to the media, companies, stock brokers and governments, among others. It allows predicting certain socio/economic aspects and taking the appropriate measures in each case before they occur. It also allows companies to improve their products in those aspects that their customers dislike most and to keep those that they like the most. These tools can even be profiled to suit certain interest groups. An example that can illustrate the previous paragraph quite well is, for example, when someone is looking for a hotel for their vacations. Normally we are guided by certain factors such as the number of stars and the opinion of the users of that hotel. The problem is that we don’t know what the person who gives the opinions is like and a father or a mother of a family may be looking for something very different from that of a young person who wants to have fun. With this trend analysis technology, a system can assess that a nearby nightclub is good for a young person but is not so interesting for other people or that the value of a stock product is going to fall imminently.
But all this technology is based on another one that analyzes people’s opinions and values the feelings expressed in them. Thus, opinion mining (also called sentiment analysis) allows us to track through the Internet and social networks the comments that people make on any subject and assess what emotion they express, if it is positive or negative and, even, the intensity.
The GPLSI Social Observer is a fully functional demo of an opinion mining system that works on Twitter tweets to assess each one of them and summarize them in a final assessment of subjectivity. In other words, Social Observer monitors the topics defined in the system and, for each tweet related to those topics, assesses whether people’s opinion is positive, negative, and to what degree.
Although the demo only works with tweets, it can be expanded to work with other sources and add more functionality, such as the aforementioned trend analysis or the system being able to learn from its mistakes to improve over time.
These features require implementation on the domain or topic of interest to the user and at GPLSI we are experts in this customization, using resources and tools for natural language processing and semantic analysis of texts in order to bring technology to people and increase the productivity of companies and their ability to react.