On Analysis and Visualization of Twitter Data

Abstract

Provision of big data analysis in a customer-friendly applicable form, with ease and affordable cost to a wide range of customers and businesses is still a big challenge for data scientists and engineers. In this study, present a framework for analysis and its visualization. We analyze Twitter messages related to a one-year span in a specific geographical area, Okinawa Island, in Japan. Our approach includes considering three types of data; basic data, further derived or extended data, and a priori information about subjects of interest. For analysis results and visualization we arrange data in a three-dimension framework of time, quality and volume. We map different elements of the data, such as number of tweets per user, time, span of time stayed in the Island, geographical location and content of messages. Based on the elements of the data within the framework, twitters are grouped and analyzed. A visual representation of analysis is presented.