Analysis of YouTube videos

OData support
Dr. Molnár Sándor
Department of Telecommunications and Media Informatics

My task is to plan and build a video analyzing environment based on the most popular video-sharing website of the world, YouTube. I have to collect an adequate amount of videos and query relevant parameters and statistics from these videos with the help of the YouTube API, that allows to access video statistics. Finally, from these videos, i have to build a database, make diagrams and evaluate the results of the analysis to get interesting information about YouTube videos.

The most important part of the task was to get acquainted with the YouTube API. With this, I could access the content that YouTube provides and use it in my own application. The most useful function of the API was the search, that helps searching videos and query some parameters, like the video ID or the statistics. I implemented it in a Java application.

While making the statistics and diagrams, I used two applications, Microsoft Excel, and RStudio that uses the programming language R. I made two types of analysis, first, I analyzed the different types of orders in searching videos, and I compared them to each other. In this case, there were more sets of videos, but the amount of videos was less. In the other case, I analyzed just one set of videos, but there were more videos in it.

The analyzed parameters of the videos were: video categories, view count, comment count, like count, dislike count and duration.

A lot of charts were made, besides averages, I made some other diagrams with the comparison of two parameters and CDF (cumulative distribution function) diagrams.


Please sign in to download the files of this thesis.