Interactive graph of which books of scripture are similar to each other

This graph depicts which books have the most similar narrative signature to each other – which was computed by scoring each verse across 264 fields, then averaging these scores for each book, and then computing the average difference in distributions by each book. To learn more about the data and methods used to generate this graph, visit Textual analysis of Book of Mormon timeline proves it could not have been faked.

For each axis, beyond being able to choose any of the specific books of scripture, one can also choose a principal component – of which principal components were calculated from a dataset where each column was one of the 264 fields measured, and each row is the average of these fields for a specific book – which therefore provides a specific value for the principal component for each book of scripture. Principal components are made up datapoints that broadly summarize all the data; for example, the first principal component basically measures whether the books have a high frequency of sermons or if they discuss more historical things. Principal components work sequentially – so once the first principal component is computed, the next principal component is designed to explain all of the parts of the data not explained by the first principal component – meaning that each principal component measures a different thing from the previous ones, and the first few principal components will be much more important than the last ones.