tertsch / okc_message_search

parse, explore, rank and search in your OKCupid message dump

This notebook:

• parses a dump of OKC messages
• compares the vocabulary size and most frquent terms in incoming and outgoing messages
• offers basic search based on tf-idf (e.g. if you search for "entropy" it returns the most relevant messages where that word appears to be important)

Getting started:

• Get your own message dump in just 5 minutes using: https://github.com/lehrblogger/OkCupid-Message-Downloader)

Coming up:

• correlating match percentage and thread length
• analysing features of entire threads with a person