Context

YouTube provide trending videos to help viewer's see what's happening on YouTube and in the world. To do this, some measures like view count, how quickly the video is generating views (hotness), views origin (including outside YouTube), the age of the videos, etc., are to be accounted.

Trending on YouTube isn't personalized and hence display the same list of videos to users in a country. The list of trending videos is updated roughly every 15 minutes in which each update, videos may move up, down, or stay in the same position in the list.

Content

This dataset is inspired by this dataset.

This dataset only contains Indonesia's trending youtube videos which are (and will be) updated daily or twice in a day. Hence, it includes not only the trending date but including the trending time. Each update may consist of 200 videos in the list.

category.json

This file consists of the category identifier number for Indonesia.

trending.csv

The dataset includes some features like video title, description, video publish time, tags, views, dislikes, etc. In general, those features are extracted from more broad properties below:

• snippet, which contains basic detail of videos
• contentDetails, which contain more detailed video information
• statistics, which contains statistics of the videos

> Repo to generate the data is available here

Acknowledgements

The dataset is extracted using the YouTube API. Heavily inspired by the work of @rsrishav and @datasnaek with similar dataset mentioned earlier.

Inspiration

Possible use cases or tasks using this dataset include, but not limited to:

• Exploratory Data Analysis
• Classification task, such as sentiment analysis, category, etc.
• Any machine learning downstream tasks