Protocol

Introduction

Following the analysis realized on the articles (see Protocol 3), we tried to understand the controversy actors behavior in a totally different channel: Youtube. The goal is to analyze their level of information about the two different queries’ meaning: Deep Web and Tor Browser. We tried to understand: first of all if Youtube could be a tool used by the actors to express their deployment within the controversy and later what were the "hot topics" discussed and supported by the protagonists of the scene.

First step

The first step was to build a solid dataset to work on. We selected the first 500 videos of each query, in order to structure a large and reliable corpus. Using Kimono and Google Refine we have acquired a large amount of data about the number of like and dislike, the number of views, number of comments, the date they were uploaded, the average rating, title and its description, the url of thumbnails and the url of the video. Using the urls, we could analize the videos and clean up our dataset from those irrelevant or those whose content was blocked or deleted. In the end of this procedure, we had a total number of 250 videos for each query. The second step was more analytical: understand how to use the available data and whether they are enough to support our goals. We immediately notice a gap between our goals and data provided, so that we decide to integrate them adding the following information:

• Video language

• Video typology.
  Examples of tipologies:Tutorial, Vlog, Amateur Documentary, Professional Documentary, Talk, Other

• Video category selected by the uploader of the video