This directory contains three primary collections of repeatable code:

1. azure_transcripts

• Python script submitted through Slurm to send audio files to Azure Cognitive Services Speech to Text service ( https://azure.microsoft.com/en-us/services/cognitive-services/speech-to-text/#features ).
  • Code includes Python streaming ( large audio files must be streamed ) and disabled profanity checks.

2. rshiny_dashboard:

• RShiny dashboard that provides:
  • Content warning click through
  • Corpus explorer tool ( https://kgjerde.github.io/corporaexplorer/ )
  • Variable pre-trained LDA topic models
    • LDA topic visualization via LDAvis ( https://github.com/cpsievert/LDAvis )
    • Association of objects with topics

3. rshiny_prep_and_demo

• prepare_conda_explorer.R
  • Merges contents of transcripts folder and metadata .xls files
  • Generates updated corpus file
• batch_topic_modeling.R
  • Generates pre-trained LDA topic models and object to topic associations in a loop
• topic_modeling.R
  • Example topic model analysis of transcripts
• corpora_explorer_demo
  • Example corporaexplorer demo used in 1.6.21 demo with instructors

4. contentwarning / corporaexplorer

• Two simple web apps that provide a content warning wrapper prior to accessing a sipmlified CorporaExplorer user interface to our corpus.