TEAM : BENCH ODYSSEY TEAM NO: 23
- PES2UG21CS148 Choppa Rasagnya
- PES2UG21CS158 Deepika Indran
- PES2UG21CS160 Devansh Guttikonda
Most popular Indian Youtube videos from June to November 2023 were scraped using Youtube's API and the following were the attributes obtained-
- Video_id
- Title
- PublishedAt
- ChannelId
- ChannelTitle
- Trending_Date
- Tags
- View Count
- Likes
- Comment Count
- Thumbnail_Link etc
The following features was used to predict the likes -
- view_count
- comment_count
- Sentiment_Scores
- time_till_trend
- Subscriber_Count
The model which we used was the XGBoost (Extreme Gradient Boosting). XGBoost is an ensemble learning technique that combines multiple decision trees to improve predictive performance.
- RMSE: 110538.93938925382
- R-squared value: 0.9597375701889096
- MAE: 31216.7363869918
- MSE: 12218857121.30113
git clone TDL_PROJECT_CODE_Team23.ipynb
Predictions need to be periodically re-evaluated as the analysis isn't based on real-time data.
-
Market Analysis: Analyze trends in video popularity, viewership, and engagement to understand audience preferences and behavior in the Indian market.
-
Thumbnail Analysis: Identify the influence of a thumbnail of a video and how it performs using CNN’s and other Deep learning techniques.
-
Sentiment Analysis: Utilize comments data to gauge audience sentiment towards specific topics or creators, aiding in reputation management and content optimization strategies.