The data is probably coming from the Census [1] through the NECS that the author mentions. If so, the documentation is here [2] (starting on page 4-42) but doesn't mention anything about Louisiana having anything out of the ordinary. It calls out Vermont, Illinois and New Jersey for having some wackiness with "pseudo districts"

[1] https://www.census.gov/geographies/mapping-files/time-series...

[2] https://www2.census.gov/geo/pdfs/maps-data/data/tiger/tgrshp...