Add steps for research download generation and deployment

[pjsier]

Eventually we want the merged data to be available on a public S3 bucket at a variety of geography levels. We'll need to add some steps for creating CSV and GeoJSON files split in a schema TBD

[pjsier]

Because running deploy_data takes a while, it could make sense that it's only triggered by the Lambda webhook set up for research uploads in the CMS

[pjsier]

This could potentially include a staging area? Also depends on the form the demographic data will take