Twitter mood light

Throughout the years, social media have been evolving and gaining worldwide popularity rapidly. It has become an important aspect for social networking and content sharing online. Twitter, an online social networking service, commands more than 650 million registered users. It is one of the fastest...

Full description

Saved in:
Bibliographic Details
Main Author: Kwa, Kah Yee
Other Authors: School of Computer Engineering
Format: Final Year Project
Language:English
Published: 2014
Subjects:
Online Access:http://hdl.handle.net/10356/59196
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Throughout the years, social media have been evolving and gaining worldwide popularity rapidly. It has become an important aspect for social networking and content sharing online. Twitter, an online social networking service, commands more than 650 million registered users. It is one of the fastest growing social networking service to date, generating millions of content that can be used to as a tell-tale of the world events. In this paper, the author demonstrates a project on a device that is able to analyse the content from Twitter and convert the content into lights to alert users of world events. Typically, the content from Twitter contains opinion and emotion of users. In the event of a natural disaster, Twitter users tend to post Tweets of negative emotions and appears in large numbers. With an abrupt increase in negative emotion Tweets, the device is capable to observe such a change and display light colours accordingly. The project would look into some of the existing text classification techniques used to classify text documents. The primary focus of the project would be to utilize an efficient text classification method that will then be able to be ported to a hardware, Arduino, which has limited resources. The author would explain in details the reason why the method was chosen and how to select the data that will be employed to construct the classifier model. This paper would conclude with the results of the text classification used and also the limitation discovered by the author during the course of the final year project. The author had also made suggestions and recommendations for future implementation of this project.