Deployment follow up tasks

Things to tidy up

• Be consistent in variable names across repos (e.g. how we refer to the data API endpoint). Note however, that Lambda has restrictions that Elastic Beanstalk doesn't, notably you can't use periods or hyphens. If we can be consistent with postgresql conventions then that would also allow us to be more consistent across environments including local environments, and would help with schema. On that note, using PGUSER would help fix a bug in the bash script in cognicity-schema (USER is defined in some shells and can't be overridden in the shell).
• We can also pass user data to EC2 instances on startup and set that up as environment variables (and we can use .env files for the rest); this would tidy up configs for stuff running on the EC2 instance and mean we don't need a separate config repo. http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/user-data.html
• There are some common patterns evolving in the way we do lambdas that sit behind the API gateway and we could move those to a separate module that we load in via packages.json.
• Add jenkins build icons ?
• Generate DOIs ? using Zenodo
• Use Travis CI more widely.
• Improve test coverage.