Is there a practical way to do this in Splunk? As far as I can tell it is basically only some sentiment analysis options which is only part of the goal.
The main goal would be figuring out if X users complained about Y.
You can achieve your main goal as well with Splunk 😉
If you have your data ready in non-binary format Splunk can index it and you can do things like "who complained when about what".....
"who complained when about what"
As example, if your data contains fields like user, productID and feedback which would hold either value of comments, reviews or complaints. In this case you can do something like this:
... | timechart count user by productID where comments=complaints
and get back a nice reports about the numbers of users who posted a complain about some product over a given time range.
In the docs about timechart search command you will see some nice example reports.
hope this helps ...