Statistical Analysis of Online News

Statistical Analysis on Online News by Vahab Pournaghshband Master of Science in Engineering Electrical Engineering and Computer Sciences University of California, Berkeley Social scientists examining questions such as media bias, or the tracking the image of people, groups, or concepts in the news, often must manually sort through the enormous amount of news data available online. StatNews aims to create a tool for processing large collections of text articles and headlines, statistically analyzing the retrieved data sets, and visualizing the results of the analysis. In particular, the idea is to come up with a new tool for extracting information from data sets with many variables. Such a tool, developed based on computational techniques, could be of interest not only to social scientists, but also to online news readers everywhere who are inundated with news items from many sources. We also believe that this will help spark a renewed interaction between applied mathematics and social sciences. “An abstract version“ of the StatNews has been developed (analysis, design, an implementation) here to smooth the road for the construction of the final, detailed version of the project. The main goal has been to figure out obstacles in advance, and well before the actual scaled project run into. Among specific questions in mind was to figure out