• information retrieval
• machine learning
• opinion mining
• sentiment analysis
• text mining

Nowadays, a huge amount of information is available from many sources and in particular from Social Media. Integrating this information into business decision making is certainly a great strategic advantage and is becoming definitely necessary. Nonetheless, the information is described in natural language, so data extrapolation is the main challenge. This work studies the main techniques of text mining and opinion mining and applies them to an enterprise application use case. The project proposes an easy-to-use system to detect events with information about place, date and kind of event gathered by websites and to extract opinions on monitored topics of the Twitter domain.
The system offers an interface to define the resource and the keyword to analyze and show the results of the extracted information.
The proposed solution makes use of text analysis, named entity recognition, topic clustering, sentiment analysis and text classification to identify the significant events that are available on the web and most influential topics that are retrievable on the social network.
The project is implemented in Python and exploits popular library such as Flask and spaCy. Under ad-hoc testing cases, the developed system has shown to be effective in the identification of events, as well as in extracting proper information from Twitter posts.