Vloggers and showmen
speak up for Net Neutrality

Introduction

We wanted to verify if the Net Neutrality theme was present on Youtube and if analyzing videos could help us going deep into the controversy. We decided to analyze video contents to study the position of authors on the subject and video comments to see how the crowd would react to the speeches. The results were compared to spot and highlight disagreements, similarities and constant trends.

As we classified videos into our categories, we noticed that 13 out of 20 placed in the vlog category, outlining how on Youtube news and knowledge spread from humble, one-man channels without big production budget. Through viral popularity and a honest, close-to-the-viewer point of view paired with a loyal and continuous effort spreading consciousness, vlogs or small productions are often more followed than TV shows or higher budget productions reversed on Youtube.

How to read the visualization

The visualization shows the most recurring concepts and keywords in Youtube under the Net Neutrality subject. The 10 most frequent words or concepts are stacked top to bottom respectively in the "speech" and in the "comment" columns.

At the initial state, every query is covered by a flat color showing its orientation: burgundy for ethical words, cerulean for technical words. A quick glance therefore shows the overall mood for original content and related comments. If hovered, the graph reveals the original words or concepts allowing a capillary examination.

How it has been done

In order to build our database, we used the "Net Neutrality" query in Youtube's search engine, and picked the 20 most viewed videos of 2014. We scraped the comments with Kimono and found the top occurrencies with Sven's TF search.

Moving on to the analysis of the original content of those videos, we downloaded subtitles using Google2srt and cleaned the formatting and timestamps obtaining pure text. Sven's TF results were used again, grouped, classified, and organized in a CSV database along comments.

When the database was complete, we moved to the representation chapter. We wanted to suggest to the user the overall polarization of ethical and technical terms and the complexity of the results at a glance, without sacrificing the opportunity of diving into specific words and sentences of specific videos.

We made a "brick wall" visualization which shows keywords as flat color bars grouped by video and, if hovered, reveals the keywords of the specific video.

Findings

As a first observation, while watching videos and manually browsing comments, we realized that on Youtube there is no proper controversy since almost everybody is pro Net Neutrality. The subject is a very discussed one though: the selected 20 videos gathered almost 18 million views and 28 thousand comments, with the John Oliver's Last Week Tonight video peaking at 7.3 million views and leading the FCC website to crash due to an excessive number of comments and connections.

Parsing the results of our analysis we realized that comments and speech keywords are usually closer to the ethical sphere. The politic section behaves differently instead: most keywords belong to the technical sphere. Among Vlogs keywords are almost balanced, since technical topics are often discussed and paired to ethical connotations.

Metadata

Timestamp: 11/12/2013 - 11/12/2014

Data source: Youtube

Related Protocol

Download data (20KB)