Introduction

Data Processing/Manipulating Actual dataset of my own YouTube Channel for practical use that YouTube Dashboard does not produce.

Datasets / Channel Information

All data is from my YouTube Channel’s live data. I have run my own YouTube Channel since April 2021, for spreading my daily life as a Graduate Student majoring in Data Analytics to my Friends, Family, and Korean subscribers. Channel Name : [KOR] 공부하는섭이 / [ENG] Studying Seob

공부하는섭이 - Jay's Channel

Libraries used

• General: Numpy, Pandas, Matplotlib
• Google API : Auth2client, Googleapiclient

YouTube Data API | Google Developers

• Korean Text Processing/Mining/Visualizing: Konlpy, Wordcloud

KoNLPy: Korean NLP in Python - KoNLPy 0.6.0 documentation

• Machine Learning: Scikit-learn

Purpose

• Extracting insights beneath the data, for enhancing the effect of propagation and channel growth.
• Excavating insight into the proper topic/time/duration/activities to upload, by extracting several indexes from data, especially focusing on the features that couldn’t be checked on YouTube Creator Dashboard.
  • Categorical Comparison
  • Manual combination of different indexes
  • Overall activities comparison at a glance
• Clustering and Visualizing the frequently mentioned words for intuitive catching of active channel viewers’(including subscribers) emotional atmosphere.

Initial Questions asked to myself

• Is there significant activities(views/comments/likes) gap among different categories?