Zovda
Most datasets treat a post as the unit and ignore what came after it. That is where the disagreement lives: corrections, counterarguments, and the sentence that changes how everyone reads the original.
A pipeline that keeps posts and drops replies is measuring what was said, not what happened.
What a reply adds
- Whether the audience accepted the claim or pushed back
- The specific objection, which is often more useful than the original
- Whether the author responded, and how
- The thread structure, which decides who saw what
The last one is the hardest to keep and the most informative about reach.
Depth is not importance
The first reply has more reach than the tenth. An argument that happens deep in a thread is real and invisible to most of the audience.
Store depth and the parent, so you can measure both the argument and its exposure.
Tone changes with depth
Replies near the top are written for everyone. Replies deep in a chain are written for the participants. Sentiment models trained on top-level replies misread the deep ones.
If you tag sentiment, tag the depth alongside it.
Deleted replies are evidence
A missing reply is often a moderation action or a retreat. Record the gap, even when the content is gone.
Zovda assembles these collections through Datuva, where each reply keeps its parent and its depth. The social media data collection shows the fields.
Tell us what you need to understand about response. We reply within one business day.