Parallelize mass conversion of transcripts to CSV

Currently, we use `thread_map` to do a lot of data fetching for getting events, sessions, transcripts, videos, etc. However our `store_transcript_as_csv` logic is just a loop that takes a long time. It would be preferable to use `process_map` (or w.e. the function name is) in a similar fashion as all the other processing we do to parallelize this process.