There is an upcoming event in the LDS Church (Mormon Church) where two leaders from the Church's hierarchy will answer questions submitted by some of the young single adults in the Church. There are about 2,600 questions submitted thus far and they are posted to a website. I wrote a Python script to scrape the questions and do text analysis on their content.
I stemmed and tokenized the words, ran them through a count vectorizer to get term counts, and then performed non-negative matrix factorization to group the texts into topics. I then performed NMF on subsets of the questions which had been filtered according to various topics or keywords.
I posted my analysis at Wheat & Tares: