Maria Ronacher

- Extracted from pdf, cleaned and anonymised data (Python) - Created multiple random forest classification models (R) - Performed text mining using different techniques: word frequency, tf-idf via Tidytext, Quanteda, tm (R) - Carried out sentiment analysis of conversations using tf-idf analysis and an external sentiment lexicon (R) - Used text mining and sentiment analysis findings to engineer additional vars which increased performance of the models (R)