I am doing a Natural language processing related project. It is a sentiment analysis task. I need to generate a dataset for the uniqueness of the work.

Is there any recommendation on how can I generate a dataset? I don't want to depend on social media websites or web scraping to generate the data.

Thank you


1 Answer 1


There are dozens of public sentiment analysis datasets you can use. Check out Kaggle datasets and Paperswithcode for that.

Another popular approach is to generate your own synthetic dataset with an LLM, either chatGPT or an open one like Llama-3. There is even a wikihow on how you can do it.

Probably a good approach to get a unique result would be to mix available datasets with a synthetic dataset.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.