Explanatory Data Analysis to Evaluate Keyword Searches for Educational Videos on YouTube with a Machine Learning Approach

Mambang Mambang,Ahmad Hidayat,Johan Wahyudi,Finki Dona Marleny
DOI: https://doi.org/10.33395/sinkron.v7i3.11502
2022-07-19
SinkrOn
Abstract:One of the most important parts of data science is the process of explanatory data analysis. This study aims to analyze learning videos on YouTube using search keywords such as learning biology, chemistry, physics, computers, mathematics, management, accounting, citizenship, history, and culture. The method used is the explanatory data analysis technique with a Machine Learning approach. The dataset used in this study uses learning video search keywords found on the YouTube digital platform. After doing a thorough analysis of all existing variables, we found that in the context of searching for learning video keywords on YouTube, the viewing variable has a heatmap correlation of 0.97 on the likes variable, 0.97 on the subscribers variable, -0.15 on the duration variable and 0.95 on the comment variable. The duration variable negatively correlates with all variables based on the analysis using a correlation heatmap using the seaborn library. Our analysis found that the number of learning videos with the search keyword Mathematics had the highest number of views among other variables. Further research can use existing variables or also add variables and add search keywords on YouTube. The data analysis approach can also be done using SPSS, R and also a Machine Learning approach with different libraries.
What problem does this paper attempt to address?