Here's an example of the metadata that I would like to include with the tweet ids. There are a few other metadata fields like "true text length" and "emoji count" that may be helpful.
You can see there are a few users with low follower counts that had very popular tweets.
Conversation
Replying to
Hopefully Twitter will allow most (if not all) of these supplemental metadata fields to accompany the actual tweet ids for the George Floyd Civil Unrest dataset.
1
4
Replying to
The amount of media data that go with this dataset is well over 25 terabytes (videos, images, etc.)
1
9
Replying to
have you been re-hydrating all/some of the tweets on your end? perhaps kept a track of the timestamp at which the tweet was collected?
I'm wondering if some researchers would be interested in the metadata based on how late/soon after the tweet was made, it was picked up by you.
1
1
Replying to
Yes. I also make sure to record the retrieval time for every social mediq object I ingest.
1