[Note: this page will be updated with new results] 

This page is dedicated to my ongoing data mining study of, in collaboration with Hannes Kettner. Erowid is a curated database of narrative reports of drug-induced experiences. My inital goal was to extract meaningful data on the subjective experience of 'drug-induced ego dissolution’. I used machine learning and natural language processing algorithms (specifically word2vec and topic modeling with LDA) to extract meaningul semantic patterns from a corpus of about 6000 narrative reports (see my 2016 presentation). More recently, I have expanded the scope of this study to investigate and compare subjective effects of psychoactive drugs in a systematical manner. Here is a visualization of an LDA model of all narratives reports for the top 20 substances of Erowid with 15 topics (also available on a separate page):