The best solution would probably be to bubble up state mutations to a single place per data model area, and factor out the http calls per data model area, group

The best solution would probably be to bubble up state mutations to a single place per data model area, and factor out the http calls per data model area, grouping as needed. I'd rather just use redux with thunks that use async/await though personally than write this from scratch.