What are the options for getting data into JS/Angular from an Impala query within a Zeppelin note?

Question

I'm currently getting data from an Impala query into Javascript/Angular, within a Zeppelin note, by using a %angular paragraph that does an AJAX call to the Zeppelin Notebook REST API (to run a paragraph synchronously, doc: https://zeppelin.apache.org/docs/0.8.2/usage/rest_api/notebook.html#run-a-paragraph-synchronously ) - this runs another paragraph in the same note, which is set to %impala and has the query. Doing this via the API means I get the data back into Javascript as the REST response to the API call. (That API link seems to be the only call in the whole API that will run a paragraph and return data from it).

However, it's been proving a bit unreliable in our big corporate setting for various reasons (network and policy related etc). My question is are there ANY other ways, within Zeppelin, that I could use to get data from a query (on a huge database), into Javascript. The data is often a few 10's of KB but could potentially by multi-MB.

For example, I know I can run pyspark .sql("My query here") type queries in a %pyspark paragraph but how do I then get the data over to JS running in the %angular para? Is the only option the same API call? I know of variable binding using the z. context but not sure that can cope with larger data sizes?

Any suggestions very welcome! Thanks