FlowJo tSNE

Are these samples compensated? What’s the effect of compensation on tSNE? Is it better to work with raw data or compensated? Thanks :)

In which cases it is necessary to modify the learning rate, perplexity and other variables? Is there a general rule for that?

How can I compare markers distribution between different samples using t-SNE?
Say I have 8 markers and 8 individuals and want to compare the distribution of the markers among them?
The same question by Juan
Do we use comp. Or uncomp? Other videos in this channel recommend to use uncomp?