Problem

To learn about sentiment analysis, I have worked on a use case that involves analyzing comments on Twitter and determining the sentiment behind them. The challenge was to create a system that could accurately analyze large amounts of data quickly and efficiently, while also presenting the results in an easy-to-understand format.

I found that traditional sentiment analysis models were not effective enough for our needs. They often failed to accurately identify the sentiment behind certain comments, leading to inaccurate results.

Check out the code in my github: https://github.com/Abdelkabir-menani/Sentiment-analysis-on-twitter-comments-NLP-

Solution

After extensive research, I decided to use the BERT model for sentiment analysis. This model is based on deep learning and has been proven to be highly accurate in identifying the sentiment behind text data.

I trained the model on the sentiment140 dataset and I developed a system that would take the comments of a given post on Twitter and run it through the BERT model for analysis. The results were then presented on a dashboard created using Dash, which allowed users to easily visualize and understand the sentiment behind each comment.

Twitter Sentiment Analysis with BERT