Computational Research of Preaching: Visualising the Viva Vox

Sermon analysis has been one of the fields of practical theology in which empirical methods have been thriving over the decades. Starting with quantitative approaches (simple word frequencies), sermon analysis is nowadays often a mix of various qualitative perspectives. The most recent developments in digital theology, however, add a new and innovative branch to the methodical forest. Computational methods range from more complex word frequencies and word embeddings, to the application of techniques from the field of natural language processing and the development of algoritms for classifying speech.
In the paper I present a proposal for this type of digital theology. Two examples test the possible contribution computational approaches to empirical theology. First, a machine learning algorithm that learns to distinghuish between preaching and prayer. Second, sentiment analysis visualise the mood of preaching.