Twitter dataset to train word embeddings

Question

I'm working on a project related to manipulating word embeddings. In order to do this, I need to train them myself on twitter data. Given Twitter's policy, I am unable to find a suitable dataset. Does anyone have one or know where I can find one?

The dataset should:

• contain public tweets
• have no specific topic, just need lots of tweets
• be pre-processed

Answer 1

Tweet Sentiment Extraction

Kaggle supports a variety of dataset publication formats.

Kaggle Twitter Datasets

You need to have an account to download the datasets. You can find the code for preprocessing in the kernels tab.