Abstract
A corpus of 1878 recorded works of historic electronic music from 1950–1999 has been collated. This novel data set empowers chronological study of variation over time, and the answering of research questions based on associated annotated metadata, such as art music versus popular music or comparing female and male composers. We describe the challenges of building our new corpus, audio analysis over all the works in it carried out via the SuperCollider Music Information Retrieval code library, and results of tackling two example research questions. The article involves some discussion of the material, but also accompanies release of the data itself.