An Analysis of Popularity of YouTube Videos Based on Upload Frequency

The aim of this thesis is to analyze a selection of a large number of observations from a dataset containing publicly available information about YouTube profiles and videos enriched with Emplifi's internal data for analytical purposes. This includes 762,525 videos released between August 9, 2021, and August 9, 2022, in the entertainment category on YouTube. The paper defines the appropriate methodology and discusses the applicability of statistical methods when working with real, big data. The data is described and subjected to exploratory analysis to get acquainted with the dataset and understand its content, performed using graphs and descriptive statistics. Subsequently, methods of analysis of variance, correlation and regression are used and will be commented on theoretically, and practically interpreted based on their real-life significance. The analysis is performed using the Python programming language. The specific focus of the thesis is on the investigation of regularity and frequency of video uploads and the impact on their success rate, in order to shed light on this aspect of the YouTube algorithm and the YouTube consumer environment. Results of the analyses do not speak in favor of a significant effect of regularity on the success rate of videos measured by their views, nor in favor of a higher success rate for a given channel's videos based purely on their quantity.