I am using twitter in my NLP research for my masters thesis. There’s something that I can’t seem to find and that is on if I can release the tweets that I’ve collected and used in my research to other researchers if they wish to improve my results with the same data?
Our terms of service do not allow for the redistribution/resyndication of Twitter content. The best way to provide your dataset for others to reuse is by providing the list of status IDs/tweet IDs that were used in the study. Another researcher who would like to do the same study would be required to then consume those tweets from the API by their own ingenuity.
What about compiling a list of all user handles? That is analogous to a list of tweet IDs.
User handles can change; user IDs (numeric) will not. So again you can share the IDs as part of your research, but as with Tweet IDs, if the users are deleted then the API will return an error. You should plan to compile and share IDs rather than user handles or Tweet content.