I'm comparing some transcript data between RefSeq and Ensembl. Ensembl seems to have many more isoforms and transcripts than RefSeq. I'm confused on exactly what the difference is because some properties between the untranslated regions and the coding regions are different between the two.
Ensembl genes are based partly on RefSeq, partly on UniProt, and partly on Havana manual annotation
so you should not expect the two resources to be equivalent.
Best to read the documentation and use the help features for each database:
If you search the Ensembl Help link for "refseq" you'll find lots of useful information. The basic difference is that RefSeq is a collection of non-redundant, curated mRNA models, whereas Ensembl is a database containing more gene models from multiple sources, mapped to the reference genome.