Did you ever end up analyzing these data? I am considering doing NLP on the experiences to produce a rating of sorts. Maybe hire some assistants to read the reviews and rate them on 0-10 scale of overall impression of niceness. One can also do cluster analysis of the experiences to see if these replicate the descriptions (e.g. deliriants should cluster together), as they should.
Did you ever end up analyzing these data? I am considering doing NLP on the experiences to produce a rating of sorts. Maybe hire some assistants to read the reviews and rate them on 0-10 scale of overall impression of niceness. One can also do cluster analysis of the experiences to see if these replicate the descriptions (e.g. deliriants should cluster together), as they should.
I see little research on these already, e.g. simple counts like https://robertorocha.info/erowid-drug-effects-visualization/ or simpler approaches like https://www.frontiersin.org/articles/10.3389/fnins.2018.00007/full